package "panelvar" in R _ error

[la-guardia-zorzin]

Please, I'm facing an error when running a PVAR model in R, using the "panelvar" package.

The error is: "Errore cholmod 'problem too large' nel file ../Core/cholmod_dense.c, linea 102"

Below, my code for the model... My base is made up of 2,618 observations, so I don't believe it's a memory problem.

varone <-pvargmm(
dependent_vars = c("Img","d_Y_Pop_c","d_Y_Pop_I", "GDP_c", "GDP_I", "d_Unemploy_c", "d_Unemploy_I"),
lags = 1,
exog_vars = c("Policy_c", "Policy_I", "W_disaster_c", "W_disaster_I"),
transformation = "fd",
data = df2,
panel_identifier = c("Time", "Country"),
steps = c("twostep"),
system_instruments = TRUE,
max_instr_dependent_vars = 99,
min_instr_dependent_vars = 2L,
collapse = FALSE
)

Thank you very much in advance for any help (or any clue, any guidance)!!!

Guidelines

• I have read and understood this category's guidelines before making this post.

[ardagurkan9]

The issue is likely caused by the order of your panel identifiers.

In panelvar, the first variable should be the panel unit and the second should be the time variable. So try changing:

panel_identifier = c("Time", "Country")

to:

panel_identifier = c("Country", "Time")

Also, your instrument setup is quite large. Using max_instr_dependent_vars = 99, collapse = FALSE, and system GMM can create a very large instrument matrix even with only 2,618 observations.

I would also try:

max_instr_dependent_vars = 4L,
collapse = TRUE

If it still fails, test the model with steps = "onestep" and system_instruments = FALSE first, then add them back one at a time.