“…The sample size n is set to 50, which is a typical value in financial literature and corresponds to roughly two months of daily data or a year of weekly data, respectively. In all considered cases we take the following parameters for the conjugate prior ν c = κ c = n, µ c = 0 k and S c = I k ; for the hierarchial prior 1 = 0.0001, 2 = 0.0001 (as in Greyserman et al (2006)), κ h = ν h = n and S h = I k ; for the informative prior δ 1 = 1 and δ 2 = 0.5, ν I = n, w I = 1/k, S I = 1.…”