CUSUM Test in ARDL Models

[themba.chirwa]

Colleagues,

I am using Eviews 9 and I have been searching if I can also use the CUSUM Test after estimation of an ARDL model. Does anyone have any suggestions on how this can be done.

Thanks in advance.

Best,

Themba Chirwa

[EViews Gareth]

Currently it isn't available, but will be via an update soon.

[Appiah-Konadu]

Currently it isn't available, but will be via an update soon.

Collegues I want to estimate the ARDL cointegration model using Eviews 7 but I cannot find it in the drop down menu. can someone help me out with the procedure

[EViews Gareth]

It was added in EViews 9.

[omerz]

Hi I am using CUSUM test for ARDL models in Eviews 11. However although I have 5 years of daily data from 2015 to 2019, my CUSUM test output graph only includes values calculated for 2019. How can I able to view CUSUM Stats for 5 years period. Thank you very much

[EViews Gareth]

More details needed:

Code: Select all

create m 2015 2019 series y=nrnd series x=nrnd equation eq1.ardl y x show eq1.rls(q) c(1) c(2) c(3)

[omerz]

For monthly data your code showed CUSUM graphics for all years. But I am working on daily data with 1826 obs. and can still only see 2019 values for CUSUM

[EViews Gareth]

Code: Select all

create d7 2015 2019 series y=nrnd series x=nrnd equation eq1.ardl y x show eq1.rls(q) c(1) c(2) c(3)

[omerz]

Your simulated data prints CUSUM test results for entire time line. However it is not the case for my equation. I have 39 dynamic terms and about 80 fixed terms in my model. Does these fixed terms could be the reason behind this problem

[EViews Gareth]

No.

[omerz]

Can you guess what is the possible problem behind this problem? I can check further specification errors in my work file

[EViews Gareth]

I'm guessing you're doing something wrong, but without further details, it is impossible to guess what.

[omerz]

As I said before I have 111 independent vars, which 39 of them enters the dynamic part of ARDL eq., in addition I have 80 fixed variables in the model. I have 1826 obs and I am using white het. resistant se estimation. My after estimating my model, I run view/stability tests - CUSUM test. This is te case... But I cant see the recursive calculation for entire sample, I can only see values for 2019