Resumo

SERRANO BAUTISTA, Ramona  e  MATA MATA, Leovardo. A conditional approach to VaR with multivariate α-stable sub-Gaussian distributions. Ens. Rev. econ. [online]. 2018, vol.37, n.1, pp.43-76. ISSN 2448-8402.

The purpose of this investigation is to propose a multivariate volatility model that takes into consideration time varying volatility and the property of the α-stable sub-Gaussian distribution to model heavy tails. The principal assumption is that returns follow a sub-Gaussian distribution, which is a particular multivariate stable distribution. The proposed GARCH model is applied to a Value at Risk (VAR) estimation of a portfolio composed by 5 companies listed in the Mexican Stock Exchange Index (IPC) and compared with the one obtained using the normal multivariate distribution, t-Student and Cauchy. In particular, we examine performances during the financial crisis of 2008.

Palavras-chave : α-stable Sub-Gaussian distribution; multivariate stable Sub-Gaussian GARCH model; Value at Risk.