Calculating CVaR and bPOE for common probability distributions with application to portfolio optimization and density estimation

DOI10.1007/S10479-019-03373-1MaRDI QIDQ2241122zbMATH OpenOpenAlexWikidataFDO

Authors Matthew Norton, Valentyn Khokhlov, Stan Uryasev

Publication date 8 November 2021

Published in Annals of Operations Research (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1811.11301

zbMATH Keywords

buffered probability of exceedanceconditional value-at-riskdensity estimationportfolio optimizationsuperquantile

Mathematics Subject Classification ID

Probability distributions: general theory (60E05) Point estimation (62F10) Nonlinear programming (90C30) Portfolio theory (91G10) Statistical methods; risk measures (91G70)

Abstract: Conditional Value-at-Risk (CVaR) and Value-at-Risk (VaR), also called the superquantile and quantile, are frequently used to characterize the tails of probability distribution's and are popular measures of risk. Buffered Probability of Exceedance (bPOE) is a recently introduced characterization of the tail which is the inverse of CVaR, much like the CDF is the inverse of the quantile. These quantities can prove very useful as the basis for a variety of risk-averse parametric engineering approaches. Their use, however, is often made difficult by the lack of well-known closed-form equations for calculating these quantities for commonly used probability distribution's. In this paper, we derive formulas for the superquantile and bPOE for a variety of common univariate probability distribution's. Besides providing a useful collection within a single reference, we use these formulas to incorporate the superquantile and bPOE into parametric procedures. In particular, we consider two: portfolio optimization and density estimation. First, when portfolio returns are assumed to follow particular distribution families, we show that finding the optimal portfolio via minimization of bPOE has advantages over superquantile minimization. We show that, given a fixed threshold, a single portfolio is the minimal bPOE portfolio for an entire class of distribution's simultaneously. Second, we apply our formulas to parametric density estimation and propose the method of superquantile's (MOS), a simple variation of the method of moment's (MM) where moment's are replaced by superquantile's at different confidence levels. With the freedom to select various combinations of confidence levels, MOS allows the user to focus the fitting procedure on different portions of the distribution, such as the tail when fitting heavy-tailed asymmetric data.

Review text

The authors of the paper are concerned with calculating superquantiles (SQ) and buffered probabilities (bPOE) for various distributions. The main result of the paper is providing closed-form formulas for various commonly used probability distributions (including normal, generalized Pareto, Student, Weibull distributions). In the cases, where finding a closed-form formula for bPOE was impossible, the authors provide a method for calculating it, by solving an equation or an optimization problem. The authors then use the derived formulas in two fields of application: portfolio theory and parameter estimation. The authors present non-classical problems of portfolio optimization, in which one of the criteria is either SQ or bPOE. They consider the possible solutions for these problems for several distributions of asset prices, that are consistent with empirical distributions. The authors formulate the problems and provides hints for solving them. In the last part they propose their own methods for estimation parameters of distribution based on superquantiles. The methods are generalizations of the method of moments and the generalized method of moments. In both cases one choses several probabilities \(\alpha_1, \ldots, \alpha_k\). In the first method one searches for values of parameters for which theoretical superquantiles equal empirical ones. In the second method one minimize a distance between theoretical and empirical quantiles. According to the authors these methods are flexible and allow to fit the tail of a distribution better.

Recommendations

• Calibrating probability distributions with convex-concave-convex functions: application to CDO pricing
• Kernel density estimation based distributionally robust mean-CVaR portfolio optimization
• Mean-CVaR portfolio selection: a nonparametric estimation framework
• A theory of the risk for empirical CVaR with application to portfolio selection
• Coherent risk measures and normal mixture distributions with applications in portfolio optimization
• Stochastic optimization problems with CVaR risk measure and their sample average approximation
• Application of Copula and Copula-CVaR in the Multivariate Portfolio Optimization
• CVaR robust mean-CVaR portfolio optimization
• Portfolio optimization with a copula-based extension of conditional value-at-risk
• Closed-form optimal portfolios of distributionally robust mean-CVaR problems with unknown mean and variance

Cites work

• Buffered Probability of Exceedance: Mathematical Properties and Optimization
• Cash flow matching with risks controlled by buffered probability of exceedance and conditional value-at-risk
• Coherent measures of risk
• Estimation and asymptotics for buffered probability of exceedance
• Fitting the generalized lambda distribution to data: a method based on percentiles
• Matching a distribution by matching quantiles estimation
• Maximization of AUC and buffered AUC in binary classification
• Random variables, monotone relations, and convex analysis
• Soft margin support vector classification as buffered probability minimization
• Tail Conditional Expectations for Elliptical Distributions
• The Cambridge dictionary of statistics.

Cited in

(25)

• A hybrid algorithm for portfolio selection: an application on the Dow Jones Index (DJI)
• Quantile-based optimal portfolio selection
• A tail measure with variable risk tolerance: application in dynamic portfolio insurance strategy
• Fitting heavy-tailed mixture models with CVaR constraints
• Bregman superquantiles. Estimation methods and applications
• Bayesian CV@R/super-quantile regression
• Ellipsoidal buffered area under the curve maximization model with variable selection in credit risk estimation
• Buffered-ranking intervals for virtual profit efficiency analysis
• Generalized PELVE and applications to risk measures
• A methodology to estimate the optimal debt ratio when asset returns, and default probability follow stochastic processes
• Buffered and reduced multidimensional distribution functions and their application in optimization
• Efficient and robust optimal design for quantile regression based on linear programming
• Globalized distributionally robust optimization problems under the moment-based framework
• Efficient global estimation of conditional-value-at-risk through stochastic kriging and extreme value theory
• Optimizing healthcare utilization using conditional value-at-risk: applications in operating theatre planning
• Ordering higher risks in Yaari's dual theory
• Censored beliefs and wishful thinking
• Algorithmic design of a conditional value-at-risk optimization model based on implied volatility for multi-period portfolio adjustment
• Risk-adaptive approaches to stochastic optimization: a survey
• Asymptotic approximations for the distribution of the product of correlated normal random variables
• Tail and memory behaviour of the cryptocurrency prices and stock market indices
• Comparing approximations and heuristics over different time horizons for a production-inventory system under the base-stock policy and conditional value-at-risk objective
• Optimal nonparametric estimation of the expected shortfall risk
• Multiperiod mean-buffered probability of exceedance in defined contribution portfolio optimization
• Conditional value-at-risk portfolio optimization in high dimensions

This page was built for publication: Calculating CVaR and bPOE for common probability distributions with application to portfolio optimization and density estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2241122)