Making Cornish-Fisher fit for risk measurement

John D Lamb, Maura E. Monville, Kai-Hong Tee

Research output: Contribution to journalArticle

24 Downloads (Pure)

Abstract

The truncated Cornish–Fisher inverse expansion is well known and has been used to approximate value-at-risk (VaR) and conditional value-at-risk (CVaR). The following are also known: the expansion is available only for a limited range of skewnesses and kurtoses, and the distribution approximation it gives is poor for larger values of skewness and kurtosis. We develop a computational method to find a unique, corrected Cornish–Fisher distribution efficiently for a wide range of skewnesses and kurtoses. We show that it has a unimodal density and a quantile function which is twice-continuously differentiable as a function of mean, variance, skewness and kurtosis. We extend the univariate distribution to a multivariate Cornish–Fisher distribution and show that it can be used together with estimation-error reduction methods to improve risk estimation. We show how to test the goodness-of-fit. We apply the Cornish–Fisher distribution to fit hedge-fund returns and estimate CVaR. We conclude that the Cornish–Fisher distribution is useful in estimating risk, especially in the multivariate case where we must deal with estimation error.
Original languageEnglish
Pages (from-to)53–81
Number of pages29
JournalJournal of Risk
Volume21
Issue number5
Early online date18 Jun 2019
DOIs
Publication statusPublished - Jun 2019

Fingerprint

Risk measurement
Skewness
Kurtosis
Conditional value at risk
Estimation error
Mean-variance
Computational methods
Value at risk
Hedge funds
Goodness of fit
Multivariate distribution
Estimation risk
Approximation
Quantile

Keywords

  • conditional value-at-risk
  • estimation error
  • goodness-of-fit
  • kurtosis
  • skewness

Cite this

Making Cornish-Fisher fit for risk measurement. / Lamb, John D; Monville, Maura E. ; Tee, Kai-Hong.

In: Journal of Risk, Vol. 21, No. 5, 06.2019, p. 53–81.

Research output: Contribution to journalArticle

Lamb, John D ; Monville, Maura E. ; Tee, Kai-Hong. / Making Cornish-Fisher fit for risk measurement. In: Journal of Risk. 2019 ; Vol. 21, No. 5. pp. 53–81.
@article{2de53256144c4962bbdecf0d8bc16c42,
title = "Making Cornish-Fisher fit for risk measurement",
abstract = "The truncated Cornish–Fisher inverse expansion is well known and has been used to approximate value-at-risk (VaR) and conditional value-at-risk (CVaR). The following are also known: the expansion is available only for a limited range of skewnesses and kurtoses, and the distribution approximation it gives is poor for larger values of skewness and kurtosis. We develop a computational method to find a unique, corrected Cornish–Fisher distribution efficiently for a wide range of skewnesses and kurtoses. We show that it has a unimodal density and a quantile function which is twice-continuously differentiable as a function of mean, variance, skewness and kurtosis. We extend the univariate distribution to a multivariate Cornish–Fisher distribution and show that it can be used together with estimation-error reduction methods to improve risk estimation. We show how to test the goodness-of-fit. We apply the Cornish–Fisher distribution to fit hedge-fund returns and estimate CVaR. We conclude that the Cornish–Fisher distribution is useful in estimating risk, especially in the multivariate case where we must deal with estimation error.",
keywords = "conditional value-at-risk, estimation error, goodness-of-fit, kurtosis, skewness",
author = "Lamb, {John D} and Monville, {Maura E.} and Kai-Hong Tee",
year = "2019",
month = "6",
doi = "10.21314/JOR.2019.408",
language = "English",
volume = "21",
pages = "53–81",
journal = "Journal of Risk",
issn = "1645-1211",
number = "5",

}

TY - JOUR

T1 - Making Cornish-Fisher fit for risk measurement

AU - Lamb, John D

AU - Monville, Maura E.

AU - Tee, Kai-Hong

PY - 2019/6

Y1 - 2019/6

N2 - The truncated Cornish–Fisher inverse expansion is well known and has been used to approximate value-at-risk (VaR) and conditional value-at-risk (CVaR). The following are also known: the expansion is available only for a limited range of skewnesses and kurtoses, and the distribution approximation it gives is poor for larger values of skewness and kurtosis. We develop a computational method to find a unique, corrected Cornish–Fisher distribution efficiently for a wide range of skewnesses and kurtoses. We show that it has a unimodal density and a quantile function which is twice-continuously differentiable as a function of mean, variance, skewness and kurtosis. We extend the univariate distribution to a multivariate Cornish–Fisher distribution and show that it can be used together with estimation-error reduction methods to improve risk estimation. We show how to test the goodness-of-fit. We apply the Cornish–Fisher distribution to fit hedge-fund returns and estimate CVaR. We conclude that the Cornish–Fisher distribution is useful in estimating risk, especially in the multivariate case where we must deal with estimation error.

AB - The truncated Cornish–Fisher inverse expansion is well known and has been used to approximate value-at-risk (VaR) and conditional value-at-risk (CVaR). The following are also known: the expansion is available only for a limited range of skewnesses and kurtoses, and the distribution approximation it gives is poor for larger values of skewness and kurtosis. We develop a computational method to find a unique, corrected Cornish–Fisher distribution efficiently for a wide range of skewnesses and kurtoses. We show that it has a unimodal density and a quantile function which is twice-continuously differentiable as a function of mean, variance, skewness and kurtosis. We extend the univariate distribution to a multivariate Cornish–Fisher distribution and show that it can be used together with estimation-error reduction methods to improve risk estimation. We show how to test the goodness-of-fit. We apply the Cornish–Fisher distribution to fit hedge-fund returns and estimate CVaR. We conclude that the Cornish–Fisher distribution is useful in estimating risk, especially in the multivariate case where we must deal with estimation error.

KW - conditional value-at-risk

KW - estimation error

KW - goodness-of-fit

KW - kurtosis

KW - skewness

U2 - 10.21314/JOR.2019.408

DO - 10.21314/JOR.2019.408

M3 - Article

VL - 21

SP - 53

EP - 81

JO - Journal of Risk

JF - Journal of Risk

SN - 1645-1211

IS - 5

ER -