I am using the rugarch package to forecast stock volatility. I want to add an external regressor and lagged versions of it if possible. The external.regressors element requires the input to be in a matrix and if I added a lagged variable it would have NA's in it, and so it would not run.
Currently, I have:
external.regressors=matrix(myts$ext_reg)
whereby myts is a time series data frame. I tried adding a one-lagged version of ext_reg into it but I am just not sure how to do that. I tried:
external.regressors=matrix(myts$ext_reg, myts$ext_reg_1)
where ext_reg_1 was produced using the Lag function from the Hmisc package. This resulted in this error:
# Error in h(simpleError(msg, call)) :
# error in evaluating the argument 'variance.model' in selecting a method for function 'ugarchspec': invalid 'nrow' value (too large or NA)
How can I format the external regressors in order to be able to be passed into external.regressors?
