Portfolio Analysis with Multivariate Normal Tempered Stable Processes and Distributions

Dirk Krause
Using the construction approach of Brownian subordination, the univariate framework of Normal Tempered Stable Lévy processes is extended to an arbitrary number of dimensions. A thorough study of the mathematical properties of the multivariate stochastic process is followed by various applications of its distributions in financial econometrics and portfolio analysis. Moreover, these distributions are employed in ARMA-GARCH models for capturing volatility clustering effects in financial markets.
doi:10.5445/ir/1000029600 fatcat:gfsoaj7airdfvcshixdtdfz6ia