Nonparametric estimation of operational value-at-risk (OpVaR)

Ainura Tursunalieva, Param Silvapulle

Research output: Contribution to journalArticleResearchpeer-review

Abstract

This paper introduces nonparametric methods for estimating 99.9% operational value-at-risk (OpVaR) and its confidence interval (CI), and demonstrates their applications to US business losses. An attractive feature of these new methods is that there is no need to estimate either the entire heavy-tailed loss distribution or the tail region of the distribution. Furthermore, we provide algorithms that facilitate applied researchers and practitioners in risk management area to implement the sophisticated empirical likelihood ratio (ELR) based methodologies to construct the CI of the true underlying 99.9% OpVaR. In a simulation study, we find that the weighted ELR (WELR) CI estimator is more reliable than the ELR CI estimator. The empirical results show that the nonparametric OpVaR estimates are consistently larger than those of other comparable methods, which provide adequate regulatory capitals, particularly during crises. The findings have implications for regulators, and effective and efficient risk financing.

Original languageEnglish
Pages (from-to)194-201
Number of pages8
JournalInsurance: Mathematics and Economics
Volume69
DOIs
Publication statusPublished - 1 Jul 2016

Keywords

  • Empirical likelihood confidence band
  • Loss severity distribution
  • Operational risk
  • Risk analysis
  • Simulation

Cite this

@article{934ccefb7bc94af8a70463df2714818f,
title = "Nonparametric estimation of operational value-at-risk (OpVaR)",
abstract = "This paper introduces nonparametric methods for estimating 99.9{\%} operational value-at-risk (OpVaR) and its confidence interval (CI), and demonstrates their applications to US business losses. An attractive feature of these new methods is that there is no need to estimate either the entire heavy-tailed loss distribution or the tail region of the distribution. Furthermore, we provide algorithms that facilitate applied researchers and practitioners in risk management area to implement the sophisticated empirical likelihood ratio (ELR) based methodologies to construct the CI of the true underlying 99.9{\%} OpVaR. In a simulation study, we find that the weighted ELR (WELR) CI estimator is more reliable than the ELR CI estimator. The empirical results show that the nonparametric OpVaR estimates are consistently larger than those of other comparable methods, which provide adequate regulatory capitals, particularly during crises. The findings have implications for regulators, and effective and efficient risk financing.",
keywords = "Empirical likelihood confidence band, Loss severity distribution, Operational risk, Risk analysis, Simulation",
author = "Ainura Tursunalieva and Param Silvapulle",
year = "2016",
month = "7",
day = "1",
doi = "10.1016/j.insmatheco.2016.05.010",
language = "English",
volume = "69",
pages = "194--201",
journal = "Insurance: Mathematics and Economics",
issn = "0167-6687",
publisher = "Elsevier",

}

Nonparametric estimation of operational value-at-risk (OpVaR). / Tursunalieva, Ainura; Silvapulle, Param.

In: Insurance: Mathematics and Economics, Vol. 69, 01.07.2016, p. 194-201.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Nonparametric estimation of operational value-at-risk (OpVaR)

AU - Tursunalieva, Ainura

AU - Silvapulle, Param

PY - 2016/7/1

Y1 - 2016/7/1

N2 - This paper introduces nonparametric methods for estimating 99.9% operational value-at-risk (OpVaR) and its confidence interval (CI), and demonstrates their applications to US business losses. An attractive feature of these new methods is that there is no need to estimate either the entire heavy-tailed loss distribution or the tail region of the distribution. Furthermore, we provide algorithms that facilitate applied researchers and practitioners in risk management area to implement the sophisticated empirical likelihood ratio (ELR) based methodologies to construct the CI of the true underlying 99.9% OpVaR. In a simulation study, we find that the weighted ELR (WELR) CI estimator is more reliable than the ELR CI estimator. The empirical results show that the nonparametric OpVaR estimates are consistently larger than those of other comparable methods, which provide adequate regulatory capitals, particularly during crises. The findings have implications for regulators, and effective and efficient risk financing.

AB - This paper introduces nonparametric methods for estimating 99.9% operational value-at-risk (OpVaR) and its confidence interval (CI), and demonstrates their applications to US business losses. An attractive feature of these new methods is that there is no need to estimate either the entire heavy-tailed loss distribution or the tail region of the distribution. Furthermore, we provide algorithms that facilitate applied researchers and practitioners in risk management area to implement the sophisticated empirical likelihood ratio (ELR) based methodologies to construct the CI of the true underlying 99.9% OpVaR. In a simulation study, we find that the weighted ELR (WELR) CI estimator is more reliable than the ELR CI estimator. The empirical results show that the nonparametric OpVaR estimates are consistently larger than those of other comparable methods, which provide adequate regulatory capitals, particularly during crises. The findings have implications for regulators, and effective and efficient risk financing.

KW - Empirical likelihood confidence band

KW - Loss severity distribution

KW - Operational risk

KW - Risk analysis

KW - Simulation

UR - http://www.scopus.com/inward/record.url?scp=84973450440&partnerID=8YFLogxK

U2 - 10.1016/j.insmatheco.2016.05.010

DO - 10.1016/j.insmatheco.2016.05.010

M3 - Article

VL - 69

SP - 194

EP - 201

JO - Insurance: Mathematics and Economics

JF - Insurance: Mathematics and Economics

SN - 0167-6687

ER -