Excessive-Dimensional Sparse Multivariate Stochastic Volatility Fashions
Authors: Benjamin Poignard, Manabu Asai
Summary: Though multivariate stochastic volatility fashions normally produce extra correct forecasts in comparison with the MGARCH fashions, their estimation methods comparable to Bayesian MCMC sometimes endure from the curse of dimensionality. We suggest a quick and environment friendly estimation method for MSV based mostly on a penalized OLS framework. Specifying the MSV mannequin as a multivariate state house mannequin, we feature out a two-step penalized process. We offer the asymptotic properties of the two-step estimator and the oracle property of the first-step estimator when the variety of parameters diverges. The performances of our methodology are illustrated by way of simulations and monetary information