BGtest for VAR model in R gives me error message

Question

I am new here, but I checked other threads and have not found solution so far. I am trying to make a VAR model in R, it looks like this:

library(vars)
library(lmtest)

# from loaded data I create model: 

v1 <- data.frame(PRG, VYS, r, GDP)
lag_order <- VARselect(v1, lag = 2, type = "const")$selection[1]

model <- VAR(v1, p = lag_order, type = "const")

Then I try bgtest for checking serial autocorrelation:

res <- data.frame(residuals(model))
test_result <- bgtest(res, order = lag_order)

but I get this error message:

Error in x$terms %||% attr(x, "terms") %||% stop("no terms component nor attribute") : no terms component nor attribute

I tried to change data type of obtained residuals but this was not helpful. I also tried to fit

VAR(v1, p = lag_order, type = "const")

directly instad of formula, but that also had no impact

Carl · Accepted Answer · 2024-04-18 16:17:15Z

VAR saves a list of lm objects in varresult.

Using some sample data from the package, model$varresult contains 4 lm objects the test may then be performed on:

library(vars)
library(lmtest)

model <- VAR(Canada, p = 2, type = "none")

bgtest(model$varresult$e)
#> 
#>  Breusch-Godfrey test for serial correlation of order up to 1
#> 
#> data:  model$varresult$e
#> LM test = 9.1337, df = 1, p-value = 0.002509
bgtest(model$varresult$prod)
#> 
#>  Breusch-Godfrey test for serial correlation of order up to 1
#> 
#> data:  model$varresult$prod
#> LM test = 0.49315, df = 1, p-value = 0.4825
bgtest(model$varresult$rw)
#> 
#>  Breusch-Godfrey test for serial correlation of order up to 1
#> 
#> data:  model$varresult$rw
#> LM test = 0.69359, df = 1, p-value = 0.4049
bgtest(model$varresult$U)
#> 
#>  Breusch-Godfrey test for serial correlation of order up to 1
#> 
#> data:  model$varresult$U
#> LM test = 0.25434, df = 1, p-value = 0.614

^{Created on 2024-04-18 with reprex v2.1.0}

Collectives™ on Stack Overflow

BGtest for VAR model in R gives me error message

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related