Using monthly data for a set of variables, we examine the out-of-sample performance of various variance/covariance models and find that no model has consistently outperformed the others. We also show that it is possible to increase the probability mass toward the tails and to match reasonably...