VAR simulation help

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

VAR simulation help

mrrox


Hi Everyone,

I am a new comer to R circle. I am trying to simulate a VAR estimation. The problem is given above in the image. Assume that the errors are iid and normally distributed.  Here the number of observations x1=x2=64.

I've written the code below. Not sure if it is correct.

pi=matrix(c(0.9,0.1,-0.1,0.7),nrow=2,ncol=2)

# errors
error1=rnorm(64,mean=0,sd=1)
error2=rnorm(64,mean=0,sd=1)

#variables x1 and x2
x1=sample(64)
x2=sample(64)

#VAR simulation#

obs = 64
set.seed(123)
E=matrix(c(error1,error2),nrow=2,ncol=64)
pi=matrix(c(0.9,0.1,-0.1,0.7),nrow=2,ncol=2)
X=matrix(c(x1,x2),nrow=2,ncol=64)
for (i in 2:obs) {
  X[,i] = pi%*%X[,i-1]+E[,i]
}

I would appreciate if you could comment whether I've correctly simulated it or not.
Thank you,
lphd