Dear Statalisters,
I've learnt a lot on this forum wrt ARDL module developed by Kripfangz and Schneider and read their 2016 Chicago Conference paper back-to-back. Although detailed, the paper left out post-estimation checks and how to interpret the results.
So, having done all need regressions, Bounds, ARDL (levels), ARDL (error correction representation) all look good but I need to ascertain the goodness of my model. I have 4 variables:
1 endogenous, 1 weakly exogenous and 2 strictly exogenous....
specified as: ardl y x1, exog(x2 x3) maxlags() aic
In summary, my post-estimation results are:
DW stat = 0.9842 (close to 1, not sure)
ARCHLM = 0.0092 (not sure, it's significant, so looks bad)
Breusch-Godfrey = 0.0002 (significant, looks bad)
Breusch-Pagan = 0.9240 (not significant, looks good)
Ramsey RESET = 0.1347 (not significant, looks good)
Multicollinearity test (mean VIF) = 2.39 (looks good)
Please how do I correct the ARCHLM and Breusch-Godfrey? Also, I tried CUSUM and Stata returned with this message:
cusum y x1, generate(y_x1)
Yvar has values outside range [0,1]....yes, because Y variable is domestic credit (% of GDP).
How can I manoeuvre this? I will appreciate your help....gracias!
I've learnt a lot on this forum wrt ARDL module developed by Kripfangz and Schneider and read their 2016 Chicago Conference paper back-to-back. Although detailed, the paper left out post-estimation checks and how to interpret the results.
So, having done all need regressions, Bounds, ARDL (levels), ARDL (error correction representation) all look good but I need to ascertain the goodness of my model. I have 4 variables:
1 endogenous, 1 weakly exogenous and 2 strictly exogenous....
specified as: ardl y x1, exog(x2 x3) maxlags() aic
In summary, my post-estimation results are:
DW stat = 0.9842 (close to 1, not sure)
ARCHLM = 0.0092 (not sure, it's significant, so looks bad)
Breusch-Godfrey = 0.0002 (significant, looks bad)
Breusch-Pagan = 0.9240 (not significant, looks good)
Ramsey RESET = 0.1347 (not significant, looks good)
Multicollinearity test (mean VIF) = 2.39 (looks good)
Please how do I correct the ARCHLM and Breusch-Godfrey? Also, I tried CUSUM and Stata returned with this message:
cusum y x1, generate(y_x1)
Yvar has values outside range [0,1]....yes, because Y variable is domestic credit (% of GDP).
How can I manoeuvre this? I will appreciate your help....gracias!
