Tail risk measures and risk allocation for the class of multivariate normal mean–variance mixture distributions

Joseph H.T. Kim, So Yeun Kim

Research output: Contribution to journalArticle


The Conditional Tail Expectation (CTE), also known as the Expected Shortfall and Tail-VaR, has received much attention as a preferred risk measure in finance and insurance applications. A related risk management exercise is to allocate the amount of the CTE computed for the aggregate or portfolio risk into individual risk units, a procedure known as the CTE allocation. In this paper we derive analytic formulas of the CTE and its allocation for the class of multivariate normal mean–variance mixture (NMVM) distributions, which is known to be extremely flexible and contains many well-known special cases as its members. We also develop the closed-form expression of the conditional tail variance (CTV) for the NMVM class, an alternative risk measure proposed in the literature to supplement the CTE by capturing the tail variability of the underlying distribution. To illustrate our findings, we focus on the multivariate Generalized Hyperbolic Distribution (GHD) family which is a popular subclass of the NMVM in connection with Lévy processes and contains some common distributions for financial modelling. In addition, we also consider the multivariate slash distribution which is not a member of GHD family but still belongs to the NMVM class. Our result is an extension of the recent contribution of Ignatieva and Landsman (2015).

Original languageEnglish
Pages (from-to)145-157
Number of pages13
JournalInsurance: Mathematics and Economics
Publication statusPublished - 2019 May 1


All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Economics and Econometrics
  • Statistics, Probability and Uncertainty

Cite this