While Conducting ARDL BOUND test I discover that my F statistics is greater than the lower bound but less than the upper bound . Upon which Pesaran and Shin 2001 concluded it to be inconclusive. Please scholar what then becomes the implication could we agree there is long run relationship or not if yes what action statisticalLy can be taken in the decision Of The right hypothesis to use Null or Alternate
you should then use the t-test of the estimated coefficient of ECM(-1)
with the ARDL the test is inconclusive if the F-statistic falls within the band, in which case it is highly recommended to go back to the traditional practice of running unit root tests and other tests for co-integration.
Try first the t-test on the coefficient of ECM(-1) -That is the second test when you PSS F-rest falls iin the inconclusive region
Tahs’s answer is incorrect - because it simply means absence of a linear combination
@ Chuck thanks please by your response if the t statistics ecm(-1) can we conclude there is a long run effect
Yes you can compare it to t-crtical values by Pss (2001) or kremers, Ericssan and Dolado (1992)
If they provided you help many times then you should recommend them unless you dont feel that way. That is the purpose of RECOMMEND net to "share" above
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series