Global VAR Modeling using the GVAR Toolbox 2.0
Individual country/region specific vector error-correcting models are estimated, where the domestic variables are related to corresponding foreign variables constructed exclusively to match the international trade pattern of the country under consideration. The individual country models are then linked in a consistent manner so that the GVAR model is solved for the world as a whole.
The GVAR methodology provides a general yet practical global modeling framework for the quantitative analysis of the relative importance of different shocks and channels of transmission mechanisms. This makes it a suitable tool for policy analysis. Its use, however, is much broader and can be used for forecasting purposes as well as other applications. It is a very general modeling framework for any large system where components are driven by weighted averages of other components - whatever the data. The approach can be equally applied to regions (e.g. regional housing markets), states or firms, to name a few possibilities.
The course will introduce the necessary tools and econometric techniques used in the development and analysis of a GVAR (Global Vector Autoregressive) model for policy analysis and forecasting. The GVAR can be briefly summarized as a two-step procedure: in the first step, small-scale country-specific models are estimated conditional on the rest of the world, represented as augmented VAR models, denoted as VARX*; in the second step, individual country VARX* models are stacked and solved simultaneously as one large global VAR model. The solution can be used for shock scenario analysis and forecasting as is usually done with standard low-dimensional VAR models.
By the end of the course the participants will have acquired detailed knowledge of and extensive hands-on experience in:
- setting up the database for such models
- specifying and estimating individual vector error correction models augmented with weakly exogenous variables (VECMX* models)
- conducting diagnostic tests for the individual model equations
- testing for weak exogeneity
- testing over-identifying restrictions on the cointegrating vectors
- including a dominant unit model in the GVAR
- solving the GVAR model using fixed/time varying weights
- performing dynamic analysis using the GVAR model including:
- persistence profiles (PPs)
- generalised impulse response functions (GIRFs)
- generalised forecast error variance decompositions (GFEVDs)
- forecasting with the GVAR
- trend/cycle decomposition of the GVAR
The course includes both theoretical and practical sessions. The theoretical sessions cover the theoretical foundations underlying the GVAR modelling approach. This knowledge provides useful background for the practical sessions. These include hands on use of the GVAR Toolbox 2.0 that will give the participants the opportunity to build their own GVAR model with step by step guidance from the instructor Dr L. Vanessa Smith.
Instructor:
The training will be delivered by Dr. L. Vanessa Smith. Dr. Smith is currently a Senior Lecturer at the University of York. Prior to that she held the position of Research Fellow and Senior Research Fellow at the Judge Business School, University of Cambridge. Her research focuses on global macroeconometric modeling and the econometric analysis of panel data models. She has been an instructor at EcoMod since 2008.