diagnostic test and stability test after ARDL estimation

[Trung BUI]

Hello users,

I estimated the following equation by using the ARDL: ln(y) = b1 * ln(x1) + b2 * ln(x2) + b3 * ln(x3) + b4 *ln(x4) + e

I am currently confused about how to perform diagnostic test and stability test after ARDL estimation.

Are there any one know how to perform those test in Stata? Please help me.

Thanks a lot in advance.

[eric_a_booth]

It should be mentioned that -ardl- (-findit ardl-) is from SSC.

Check out -cusum6- from SSC for examining stability, and -help regress_postestimationts- for a list of post estimation checks for 1st order & higher order autocorrelation, heteroscedasticity, etc. If you are looking for some other tests we'd need more information (name/type, citation, etc).

- Eric

[Sebastian Kripfganz]

With the ardl command you can use the regstore() option if you want to apply standard regress postestimation commands. Please see help ardl for details about this option. Also, the following Statalist topic might be of interest to you:
ARDL in Stata

[Trung BUI]

Hello Eric,

How could I perform Chow predictive failure test after ARDL estimation?

Thanks a lot in advance.

[Marco Giansoldati]

Dear Members,
I would like to test the presence of one or multiple structural breaks on the macroeconomic series I using as explanatory variables of imports. My investigation covers only three countries, which I selected from a database which include a much larger number of countries.
In order to select the countries of interest I use the if option and I employ the ardl command. An example of my specification is as follows:

ardl mgsv rmp nc itv xgsv if ifscode==939, ec1 lags(1 5 1 1 1) regstore(ardl9)

My intention is to employ cusum6 command as follows

cusum6 mgsv rmp nc itv xgsv if ifscode==939, cs(cusum) lw(lower) uw(upper)

but I get the following error message:

command may not be used with panel data
r(459);

I understand from the help of the cusum6 help that it can be employed only with time series data, but I was wondering if there is a way to employ the command for separate countries.
I do hope that the question is not too silly and I do thank everyone for the kind attention.
Marco