A1343
Title: Tail dependence in Gram-Charlier type multivariate distributions: The relevance of the moment spillovers
Authors: Javier Perote - University of Salamanca (Spain) [presenting]
Andres Mora Valencia - Universidad de los Andes (Colombia)
Ines Jimenez - University of Salamanca (Spain)
Abstract: The purpose is to introduce the moment interaction between different assets in the semi-nonparametric modeling of the multivariate distribution. The traditional multivariate Gram-Charlier distribution is extended by incorporating the crossed products of Hermite polynomials in the multivariate expansion. The weights of these additional terms convey spillover effects between different moments and help understand tail dependence. Bivariate portfolios are analyzed where skewness and kurtosis may interact between them and across different assets in the portfolio, showing that these new parameters may be significant pieces of information, particularly for risk measuring. Model performance for risk assessment is tested with backtesting techniques considering equally weighted portfolios of S\&P 500 and Nasdaq 100 indices and major cryptocurrencies (Bitcoin and Ethereum), the latter with high-frequency data. Results show an adequate performance of the expanded Gram-Charlier multivariate distribution in terms of value-at-risk and medium shortfall, especially for high confidence levels.
