ARDL IN E VIEWS 9 AND 10

[ibrahimtr]

dear Mr Gareth, kindly explain to me the reason behind the differences in results generated by e views 9 and 10 using the same data. having tried on several occasions I realized that the coefficients are the same but the differences emanated from standard errors. the standard errors in e views 10 are far lower than that of e views 10 leading to different inferences. which one is more correct and how can we justify it?

[EViews Gareth]

Those releases are quite old, so not sure I can remember a specific change.

Given that ARDL is a simple least squares, you can verify the results by running an LS manually.