A Bayesian semiparametric vector Multiplicative Error Model

Interactions among multiple time series of positive random variables are crucial in diverse financial applications, from spillover effects to volatility interdependence. A popular model this setting is the vector Multiplicative Error Model (vMEM) which poses a linear iterative structure on dynamics conditional mean, perturbed by multiplicative innovation term. main limitation vMEM however its restrictive assumption distribution Bayesian semiparametric approach that models as an infinite location-scale mixture multidimensional kernels with support orthant used address major shortcoming vMEM. Computational complications arising constraints avoided through formulation slice sampler parameter-extended unconstrained version model. The method applied simulated and real data flexible specification obtained outperforms classical ones terms fitting predictive power.