Performance of vector autoregressive models influenced by collinearity and autocorrelated error
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Department of Statistics, Nasarawa Sate University Keffi.
DOI
Abstract
The goal of Vector Autoregression (VAR) or Bayesian Vector Autoregression (BVAR) is the characterization of the dynamics and endogenous relationships among time series. The VAR models are known for their applications to forecasting and policy analysis. This paper compares the performance of VAR and Sims-Zha Bayesian VAR models when the multiple time series are jointly influenced by different levels of collinearity and autocorrelation in the short term. Simulations on different levels of collinearity and autocorrelation were performed (viz. -0.9, -0.5, 0, +0.5, +0.9). The results revealed that the VAR(2) model dominated for levels (-0.5, 0, +0.5) of autocorrelation irrespective of the collinearity level, except when T=16. The BVAR model dominated for levels (-0.9, +0.9) of autocorrelation regardless of the collinearity level, except when T=128. The forecasting models were found to depend on the Time Series data structure and the time series length.
Description
Citation
Adenomon, M.O. & Oyejola, B.A. (2019) Performance of vector autoregressive models influenced by collinearity and autocorrelated error
