Dear friends,

I am using WDI dataset which is time series for India only. I ran first DF test and PP test for stationarity where I found all the VRs I(1). Further, I checked for any co-integration using bounds test and the result was in favour of one co-integrating eqn. The same result came using Johansen test.

The ARDL test gives satisfactory results and favoured long run relationship but when I ran vecm on the same for short run I saw ECT is positive and non-significant. Even the coefficients direction had been changed from the original ARDL.

Further granger test suggests there is no unidirectional or bidirectional causality running from IVs to DVs.

Please suggest remedial procedure. What can be done in this regard? Is it necessary to show VECM even after using ARDL or ARDL is enough?

Thank you in advance.

Suraj Sharma, PhD

More Suraj Sharma's questions See All
Similar questions and discussions