Bayesowska estymacja modelu lokalnego poziomu o rozkładach dopuszczających warunkowy rozkład t-Studenta i zmienną wariancję
Bayesian Estimation of Local Level Model with Student-t Disturbances and Time-Varying Conditional Variance

Summary/Abstract: The paper presents the general LL (Local Level) model with time-varying conditional variance, recently proposed by Stock and Watson. The main purpose is to present the Bayesian estimation and model comparison of different local level models with Normal GARCH, Student-t GARCH and SV disturbances. We are particularly interested how the different specifications of the conditional variance affect the explanatory power of a set of competing models. We apply the LL models to logarithmic transformations of the original prices of Żywiec, Polish company listed on the WSE. The model selection and posterior estimates provide strong evidence in favor of a model with SV disturbances in the core component, and the transitory component.

• Details
• Contents