Πλοήγηση ανά Συγγραφέα "Gardikiotis, Nikolaos"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

Identification of macroeconomic shocks on large var models
(2022-03-11) Gardikiotis, Nikolaos; Γαρδικιώτης, Νικόλαος; Athens University of Economics and Business, Department of Economics; Tzavalis, Elias; Varthalitis, Petros; Dendramis, Yiannis
The goal of this dissertation is to study large vector autoregressive models. The vector autoregression has been used extensively to describe the behavior of multivariate time series across time. In recent years, researchers use many economic variables in vector autoregressions as there is evidence that large vector autoregressions tend to forecast better. However, there are some challenges in estimating large vector autoregressions. Typically, large vector autoregressions have far more parameters than observations and some appropriate methods, like shrinkage, regularization or dimension reduction must be used to deal with parameter uncertainty. The estimation of large vector autoregressions also involves the manipulation of large matrices which is computationally intensive. If we allow for more flexible covariance structures in the large vector autoregression, the challenges grow even bigger. We will review a couple of vector autoregressive models, such as Bayesian vector autoregressions, Time-Varying parameter vector autoregressions and the Factor-Augmented vector autoregression (FAVAR). To deal with the “curse of dimensionality”, we will augment the vector autoregression with factors. The FAVAR will contain 123 U.S. macroeconomic indicators and will be estimated by a 2-Step principal components approach. Our results will be compared to those by Bernanke, Boivin and Eliasz from their 2005 paper. Examining the impulse responses of some key macroeconomic indicators to a monetary shock, we find that the FAVAR is able to identify the monetary transmission mechanism.