ARDL and OLS

[luqman safdar]

Hi
any one cal tell me that , if variable are not integrated at same i-e 1(0) and 1(1) order but they integrated at order 1(1)... can i allow to run multiple regression (OLS) to measure the relationship of dependent and independent variables... and for estimating ling run then further move towards Bound testing approach..

i-e first use OLS and then ARDL.....is it possible ?? or run only ARDL??

[umair khan]

if your data is stationary at different levels e.g some variables are on level form and some are at first difference, then u cannot apply ols u should go for the ardl approach

[luqman safdar]

thnx
one more question...
in estimating Long run model after bound testing.. it not compulsory to use lagged values of independent variables??? we can run with out taking lag values...only for estimating short run we must have to put D or Difference operator before our dependent and independent variables???

Am i right...????? this called ARDL (1,0000) model