Tail behavior of sums and differences of log-normal random variables

DOI10.3150/14-BEJ665MaRDI QIDQ5963508zbMATH OpenFDO

Authors Archil Gulisashvili, Peter Tankov

Publication date 22 February 2016

Published in Bernoulli (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1309.3057, https://projecteuclid.org/euclid.bj/1443620857

zbMATH Keywords

importance samplingLaplace's methodMonte Carlo methodmultidimensional Black-Scholes modelmultidimensional log-normal distributionrisk managementstress testingtail behavior

Mathematics Subject Classification ID

Large deviations (60F10) Limit theorems in probability theory (60F99) Monte Carlo methods (65C05) Financial applications of other theories (91G80)

Abstract: We present sharp tail asymptotics for the density and the distribution function of linear combinations of correlated log-normal random variables, that is, exponentials of components of a correlated Gaussian vector. The asymptotic behavior turns out to depend on the correlation between the components, and the explicit solution is found by solving a tractable quadratic optimization problem. These results can be used either to approximate the probability of tail events directly, or to construct variance reduction procedures to estimate these probabilities by Monte Carlo methods. In particular, we propose an efficient importance sampling estimator for the left tail of the distribution function of the sum of log-normal variables. As a corollary of the tail asymptotics, we compute the asymptotics of the conditional law of a Gaussian random vector given a linear combination of exponentials of its components. In risk management applications, this finding can be used for the systematic construction of stress tests, which the financial institutions are required to conduct by the regulators. We also characterize the asymptotic behavior of the Value at Risk for log-normal portfolios in the case where the confidence level tends to one.

Review text

Motivated by problems in risk management, this paper addresses asymptotic approximations for the tail behaviour of sums and differences of log-normal random variables. Section 2 focuses on the left tail of a sum of correlated log-normal random variables. Asymptotic formulas for the distribution function and density are developed under mild conditions on the covariance structure. These results complement those developed for the right tail by \textit{S. Asmussen} and \textit{L. Rojas-Nandayapa} [Stat. Probab. Lett. 78, No. 16, 2709--2714 (2008; Zbl 1151.60009)] and indicate a very different type of behaviour where, for the left tail, the correlation structure is important. In Section 3, a right tail approximation for the difference of two log-normal sums is derived. The approximations can be used directly or to construct variance reduction procedures to estimate the tail probabilities via Monte Carlo techniques. In Section 4, an efficient importance sampling technique is described for the evaluation of the tail event probabilities via Monte Carlo methods. Section 5 applies the approximations to risk management issues, such as stress tests, developed in the context of the multidimensional Black-Scholes model. A sharp asymptotic formula is developed for the value at risk when the confidence level tends to one.

Recommendations

• Asymptotic behavior of tail density for sum of correlated lognormal variables
• Efficient simulation of tail probabilities of sums of correlated lognormals
• On the efficient simulation of the left-tail of the sum of correlated log-normal variates
• On the tail probabilities of aggregated lognormal random fields with small noise
• Fast and accurate computation of the distribution of sums of dependent log-normals
• On log-normal convolutions: an analytical-numerical method with applications to economic capital determination
• Second order asymptotics of aggregated log-elliptical risk
• Exponential family techniques for the lognormal left tail
• scientific article; zbMATH DE number 219297
• On Sums of Lognormal Random Variables

Cites work

• An Introduction to Heavy-Tailed and Subexponential Distributions
• Application of large deviation methods to the pricing of index options in finance.
• Asymptotic behavior of tail density for sum of correlated lognormal variables
• Asymptotic results for tail probabilities of sums of dependent and heavy-tailed random variables
• Asymptotic results for the sum of dependent non-identically distributed random variables
• Asymptotic tail probabilities of sums of dependent subexponential random variables
• Asymptotics beats Monte Carlo: the case of correlated local vol baskets
• Asymptotics of sums of lognormal random variables with Gaussian copula
• Bounds for the sum of dependent risks having overlapping marginals
• Effects of Poiseuille flow on peristaltic transport
• Efficient simulation of tail probabilities of sums of correlated lognormals
• Gaussian multiplicative chaos and applications: a review
• scientific article; zbMATH DE number 3133146 (Why is no real title available?)
• scientific article; zbMATH DE number 3960673 (Why is no real title available?)
• scientific article; zbMATH DE number 193078 (Why is no real title available?)
• scientific article; zbMATH DE number 3512673 (Why is no real title available?)
• scientific article; zbMATH DE number 3338262 (Why is no real title available?)
• On sums of conditionally independent subexponential random variables
• On Sums of Lognormal Random Variables
• Optimal approximations for risk measures of sums of lognormals based on conditional expectations
• Pricing and Hedging Spread Options
• Stochastic simulation: Algorithms and analysis
• Sums of Dependent Nonnegative Random Variables with Subexponential Tails
• Tail asymptotics for dependent subexponential differences
• Tail asymptotics for the sum of two heavy-tailed dependent risks
• The log-normal approximation in financial and other computations
• The sum and difference of two lognormal random variables

Cited in

(30)

• Asymptotic behavior of tail density for sum of correlated lognormal variables
• Asymptotic behaviour of high Gaussian minima
• On the efficient simulation of the left-tail of the sum of correlated log-normal variates
• The sum and difference of two lognormal random variables
• Efficient importance sampling for large sums of independent and identically distributed random variables
• On density functions related to discrete time maximum of some one-dimensional diffusion processes
• Efficient algorithms for tail probabilities of exchangeable lognormal sums
• High minima of non-smooth Gaussian processes
• Fast and accurate computation of the distribution of sums of dependent log-normals
• On log-normal convolutions: an analytical-numerical method with applications to economic capital determination
• Practical aspects of the aggregation of two risks in the Solvency II standard formula
• Tail asymptotics of random sum and maximum of log-normal risks
• Second order asymptotics of aggregated log-elliptical risk
• Discrete sums of geometric Brownian motions, annuities and Asian options
• On the tail probabilities of aggregated lognormal random fields with small noise
• Exponential family techniques for the lognormal left tail
• Comparing approximations for risk measures of sums of nonindependent lognormal random variables
• Tail behavior of the Laplace and logistic distributions
• Importance Sampling for Sums of Lognormal Distributions with Applications to Operational Risk
• Implied volatility of basket options at extreme strikes
• On the probability density function of baskets
• The parametric and additive partial linear regressions based on the generalized odd log-logistic log-normal distribution
• Approximating the Laplace transform of the sum of dependent lognormals
• Aggregation of log-linear risks
• A note on portfolios of averages of lognormal variables
• A fast and accurate numerical method for the left tail of sums of independent random variables
• Efficient simulation of tail probabilities of sums of correlated lognormals
• Tail-adaptive Bayesian shrinkage
• Optimal approximations for risk measures of sums of lognormals based on conditional expectations
• Asymptotics of sums of lognormal random variables with Gaussian copula

This page was built for publication: Tail behavior of sums and differences of log-normal random variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5963508)