Abstract

ON THE STUDY OF THE BASIC PROPERTIES OF MULTIVARIATE AUTOREGRESSIVE MODELS

Usoro, Anthony E.

---

Models containing numerous response and predictive time variables, known as multivariate autoregressive models, establish a link between each response and the lag terms of both the response and predictive variables. Multivariate time series models, like univariate time series models, contain some basic properties that distinguish each model. The basic features of Multivariate Autoregressive Models (MARM), also known as Vector Autoregressive Models (VARM), are investigated in this study. The work focuses on using model parameter estimations to derive the variance, autocorrelations, and crossautocorrelations of multi-dimensional VAR models. The features of broad VAR models, such as variances, autocovariances, cross-autocovariances, autocorrelations, and crossautocorrelations, are calculated and validated using empirical examples. Keywords: Vector Autoregressive Models, Autocovariance/Cross-Autocovariance and Autocorrelation/Cross-autocorrelations

PDF