ARDL optimal lag selection

[oussama zen]

Dear community,
I need your help.
I want to calculate optimal lag selection for an ARDL model, when I use the VARSOC command, I receive the following message in return:

unable to allocate matrix;
You have attempted to create a matrix with too many rows or columns or attempted to fit a model
with too many variables.

You are using Stata/IC which supports matrices with up to 800 rows or columns. See limits for
how many more rows and columns Stata/SE and Stata/MP can support.

If you are using factor variables and included an interaction that has lots of missing cells,
try set emptycells drop to reduce the required matrix size; see help set emptycells.

If you are using factor variables, you might have accidentally treated a continuous variable as
a categorical, resulting in lots of categories. Use the c. operator on such variables.
r(915);

has anyone had this problem before.

By advance thank you.
all the best

[Sebastian Kripfganz]

You do not need to use the varsoc command for the optimal lag selection in an ARDL model. If you are using the ardl command, it automatically does it; see the ardl help file for the options aic or bic.

[oussama zen]

thank you Mr. sebastien for this answer.
indeed, when I use the ardl command with aic and for example I set maxlag to (4), I get the following message:

note: lnaiss_as omitted because of collinearity
note: L.lnaiss_as omitted because of collinearity
note: L2.lnaiss_as omitted because of collinearity
note: L3.lnaiss_as omitted because of collinearity
note: L4.lnaiss_as omitted because of collinearity
note: lgros_edu omitted because of collinearity
note: L.lgros_edu omitted because of collinearity
note: L2.lgros_edu omitted because of collinearity
note: L3.lgros_edu omitted because of collinearity
note: L4.lgros_edu omitted because of collinearity
Collinear variables detected.

furthermore, when i set maxlag (3) i get the following message:

note: L3.lgros_edu omitted because of collinearity
Collinear variables detected.

while, i put maxlag(2), my algorithme runs well.

can i consider that my optimal lag equal to 2?

all the best

[Sebastian Kripfganz]

It might be that you only have a small sample size. You cannot estimate large models in that case and may have to either limit the number of independent variables or their maximum lag order.

[oussama zen]

Thank you mr Sebastian,
you are right.
i have small sample, and for this raison Ilthe ardl model has been chosen.
warmest regards