Changelog
Listed below are all main changes of the GVAR Toolbox 1.0 since its release in December 2010.
GVAR Toolbox 2.0 (released August 2014)
The following items were fixed and/or added:
fixed a problem when changing the sample size for estimation under time-varying weights
fixed a problem when using annual data related to the printing format of the annual dates
fixed a problem with the implementation of the weak exogeneity test for lag orders p*=q*. A separate problem was also amended for the case where p*,q*>2.
fixed a problem in relation to the construction of the trade share matrix for a single year
fixed a problem with the bootstrap median estimates and bootstrap bands under SGIRFs and SGFEVD (this problem also affected the persistence profiles computed when SGIRFs were selected)
made minor changes to the bootstrap procedure which will have a minor effect on the bootstrap results
increased the size of the user-specified weighting matrix previously set to a maximum of 49.
GVAR Toolbox 1.1 (released July 2011)
The following items were fixed and/or added:
updated the real GDP series for China to reflect the latest revisions in the data
enabled the overidentifying restrictions function for imposing restrictions on the cointegrating vectors
improved the Matlab routines for faster printing of the output
fixed a problem related to the impulse responses where negative and positive global shocks gave the same result
fixed a problem related to the case where no global variables are included in the GVAR.