R: Use VAR model to predict response to change in values of certain variables

Hi Mat

Yes, this is exactly what I would like to do. Only problem this that I
generate the model using VECM and and then convert to VAR form using
vec2var. (i.e. the object is of type vec2var, not varest)

I am not sure how to extract the lm from vec2var.


Karl


On Mon, Nov 23, 2009 at 3:46 PM, Matthieu Stigler <
matthieu.stigler at gmail.com> wrote:

Hi Karl

Please provide reproducible code when you ask questions, we don't have
access to your local drive!!
So use rather datasets in R, as for example:
data(Canada)
colnames(Canada)
?var.2c <- VAR(Canada, p = 1, type = "none")

You will note that there is a predict.varest function which does forecasts:
? ? pred<-predict(var.2c, n.ahead = 8, ci = 0.95)

This is not exactly what you want... you want to change some
forecasted values, and see the influence, right? This is not available
i the predict.varest. You will need to do some manipulations,
basically extracting the lm object and using predict(lm, newdata)


Construct data frame with forecasted values. It should include all
variables you see when in the output of var.2c. To keep simple, I used
the simplest model (1 lag, no intercept), so need only to give 4
values (the lag-1). For the example, I took last dat of data Canada,
change as you want:
:
yourPred<-as.data.frame(matrix(Canada[nrow(Canada),], nrow=1))
colnames(yourPred)<-colnames(model.frame(var.2c$varresult$e))[-1] #you
need to give same names as in var result!

predict(var.2c$varresult$e, newdata=yourPred)

Note that this is exactly the same as what does forecast.varest,
except you can now change it manually!

pred$fcst$e[1,1]

And now you can change the values in yourPred as you want.

Hope this helps

Mat

2009/11/23 Karl Schriek <kschriek at gmail.com>:

Hi

I've fitted a VECM model in R, and converted in to a VAR representation.

I

would like to use this model to predict the future value of a response
variable based on different scenarios for the explanatory variables.

Here is the code for the model:

library(urca)
library(vars)

input <-read.csv("data.csv")
ts <- ts(input[16:52,],c(2000,1),frequency=4)
dat1 <- cbind(ts[,"dx"], ts[,"u"],

ts[,"cci"],ts[,"bci"],ts[,"cpi"],ts[,"gdp"])

args('ca.jo')
vecm <- ca.jo(dat1, type = 'trace', K = 2, season =
NULL,spec="longrun",dumvar=NULL)
vecm.var <- vec2var(vecm,r=2)

Now what I would like do is to predict "dx" into the future by varying

the

others. I am not sure if something like "predict dx if
u=30,cpi=15,bci=50,gdp=..." in the next period would work. So what I have

in

mind is something along the lines of: increase "u" by 15% in the next

period

(which would obviously impact on all the other variables as well,

including

"dx") and predict the impact into the future.

Also, I am not sure if the "vec2var" step is necessary, so please ignore

it

if you think it is redundant.

Thanks
Karl

? ? ? ?[[alternative HTML version deleted]]

? ? ? ?[[alternative HTML version deleted]]