fitted values Vars package in R

Question

I am using the VARS package in R and the dataset Canada found in the package itself. I am not understanding the reason why the fitted model has 2 data points less than the actual time series Canada. The Canada time series is of size 84*4 whereas the fitted are of size 82*4. I am using the following code:

library(vars)
        rm(list=ls(all=TRUE))
        data(Canada)
        #to plot the time series
        par(mfrow=c(1, 1)) 
        plot(Canada, plot.type="m", mar=c(gap=0.3, 5.1, gap=0.3, 2.1))
        #VARselect used to get the order of the time series 
        VARselect(Canada, lag.max = 5, type="const")
        #VAR(2) estimation
        var.2c <- VAR(Canada, p = 2, type = "const")
        fitted(var.2c)

Answer 1

The parameter p is for the lag order, and if p = 2 , and the time series has 84 observations, then the observations used in VAR must be 84 - 2 = 82

You can change p and see what happens:

var.2c <- VAR(Canada, p = 1, type = "const")
fitted(var.2c)

var.2c$obs
# 83

https://www.rdocumentation.org/packages/vars/versions/1.5-2/topics/VAR