Vector Autoregressive Models


This chapter reviews vector autoregressive (VAR) modelling and its applications in tourism demand research. Since the 1980s, VAR models have been a popular tool in macroeconomic analysis, in which endogeneity is of particular concern. This chapter revisits the classic VAR model. Then it introduces a recent advancement called the global VAR (GVAR) model, which is well suited to modelling large high-dimensional systems with multiple cross-sections. In addition, this chapter touches on the Bayesian approaches to VAR modelling. In the context of tourism demand research, VAR models can be used to capture the interrelations between tourism variables and economic variables and to simulate impulse responses to economic shocks. Using global tourism demand data for 24 major economies, this chapter demonstrates the applications of the classic VAR, GVAR, Bayesian VAR (BVAR) and Bayesian GVAR (BGVAR) models and compares their forecast accuracy with the accuracy of commonly used univariate time series models.

Publication DOI:
Divisions: College of Business and Social Sciences > Aston Business School > Centre for Personal Financial Wellbeing
College of Business and Social Sciences > Aston Business School > Economics, Finance & Entrepreneurship
Additional Information: This is an Accepted Manuscript of a book chapter published by Routledge/CRC Press in Econometric Modelling and Forecasting of Tourism Demand: Methods and Applications on [27/10/2022], available online:
Uncontrolled Keywords: Vector autoregressive models,Global VAR,Bayesian VAR,Forecasting,Tourism demand
ISBN: 9781032216423, 9781032216416, 9781003269366
Last Modified: 08 Dec 2023 12:59
Date Deposited: 12 Oct 2022 09:02
Full Text Link:
Related URLs: https://www.rou ... k/9781032216416 (Publisher URL)
PURE Output Type: Chapter
Published Date: 2022-10-27
Authors: Cao, Zheng (ORCID Profile 0000-0002-3545-7313)



Version: Accepted Version

Access Restriction: Restricted to Repository staff only until 27 June 2024.

Export / Share Citation


Additional statistics for this record