Description Usage Arguments Details Value Author(s) References See Also Examples
GVAR
computes VECMs for all regions and stacks the models to a Global Vector Autoregressive Model
1 2 |
Data |
timeseries data as list (each entry is a matrix of a subsystem of variables, if |
tw |
time window, vector of start and end point, if |
p |
scalar/vector of endogenous lags, if a scalar is provided the same lag length is used for all subsystems |
q |
scalar/vector of (weakly) exogeneous lags, if a scalar is provided the same lag length is used for all subsystems |
r |
scalar/vector of cointegrating relations, if a scalar is provided the same cointegration rank is used for all subsystems, if set to |
weight |
weight matrix, the diagonal elements need to be zero |
Case |
single value/vector of cases ( |
exo.var |
if |
d |
list showing which strictly exogeneous variables enter the subsystem equations, if |
lex |
scalar/vector of strictly exogeneous lags, if a scalar is provided the same lag length is used for all subsystems |
endo |
list of endogenous variables used in each subsystem, if |
ord |
vector used if variables in the different subsystem don't appear in the same order, order of each subsystem is concatenated to one vector, if |
we |
list with numbers of weakly exogeneous variables included in each VECM, corresponds to numbers in |
method |
select cointegrating rank by max. eigenvalue ( |
caseTest |
provide test statistics regarding the intercept/trend structure |
weTest |
perform F test for weak exogenity |
The function computes a VECM for every subsystem before stacking the results to a GVAR model.
Specification of input here.
An object of class GVAR
containing the following items:
subsys |
subsystem names |
Data |
data |
we.vecms |
VECMs of the subsystems |
X |
data as one single matrix |
bigT |
length of time series data |
r |
vector of cointegration ranks of the VECMs |
Case |
vector of intercept/trend behaviour of the VECMs |
W |
multiplier matrix to generate endogenous and weakly exogenous variables from X |
G |
multiplier matrix for the current variables |
H |
multiplier matrix for the lagged variables |
Upsilon.0 |
multiplier matrix for the current strictly exogenous variables |
Upsilon |
multiplier matrix for the lagged strictly exogenous variables |
c.0 |
multiplier matrix for the intercept |
c.1 |
multiplier matrix for the trend |
caseTest |
test statistics for case selection |
weight |
weight matrix used to calculate the weakly exogenous variables |
U |
residuals of the GVAR |
U.cov |
residual covariance matrix |
arguments |
arguments passed to GVAR function, including lags, variable types,... |
Martin Summer, Klaus Rheinberger, Rainer Puhr
Stephane Dees, Filippo di Mauro, Hashem Pesaran, and L. Vanessa Smith. Exploring the international linkages of the Euro area: A global VAR analysis. Journal of applied Econometrics, 22(1), 2007.
Soeren Johansen. Likelihood-Based Inference in Cointegrated Vector Auto-Regressive Models. Advanced Texts in Econometrics. Oxford University Press, 1995.
M. Hashem Pesaran, Yongcheol Shin, and Richard J. Smith. Structural analysis of vector error correction models with exogenous I(1) variables. Journal of Econometrics, 97:293-343, 2000.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 | data(pesaran26)
c.names <- names(Data)[-length(Data)]
p <- c(2,2,2,1,2,2,1,2,2,2,2,1,2,1,1,2,2,2,2,2,2,1,2,2,2,2)
q <- c(2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)
lex <- q
endo <- ord <- we <- d <- vector("list",length=length(c.names))
names(endo) <- names(ord) <- names(we) <- names(d) <- c.names
# base country: usa
endo[[1]] <- c(1:3,5:7)
ord[[1]] <- c(1:3,5:7)
we[[1]] <- c(1:2,4)
d[[1]] <- NULL
# countries with 6 endogenous variables:
for (j in c("EuroArea", "Japan", "UK", "Sweden", "Switzerland", "Norway", "Australia", "Canada", "NewZealand", "Korea", "Safrica"))
{i <- which(c.names==j); endo[[i]] <- ord[[i]] <- 1:6}
# countries with 5 endogenous variables:
for (j in c("Argentina", "Chile", "Malaysia", "Philippines", "Singapore", "Thailand", "India"))
{i <- which(c.names==j); endo[[i]] <- ord[[i]] <- 1:5}
# countries with 4 endogenous variables:
for (j in c("China", "Brazil", "Mexico", "Peru", "Indonesia", "Turkey"))
{i <- which(c.names==j); endo[[i]] <- ord[[i]] <- c(1:2,4:5)}
# Saudi Arabia
endo[[21]] <- ord[[21]] <- c(1:2,4)
# all countries but us
for (i in 2:length(we))
{
we[[i]] <- c(1:3,5,6)
d[[i]] <- 1
}
Case <- "IV"
r <- c(2,1,1,4,3,3,3,2,2,1,2,3,3,4,4,3,3,4,1,2,3,3,2,1,1,1)
res.GVAR <- GVAR(Data=Data,r=r,p=p,q=q,weight=weight,Case=Case,exo.var=TRUE,d=d,lex=lex,ord=ord,we=we,endo=endo,method="max.eigen")
# view vecm models
res.GVAR$we.vecms
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.