Since you are using ARDL and e- views apply Bounds F test to see if the cointegration persists or not, It is possible with e-views version 9. This F test will tell if we can proceed further or not

If the F bounds test value is not bigger than any of the bound values then there is no cointegration among these variables. If the F test value is small than the bound values then you have to change the variables by adding trend variable. If the F test value is larger then you can go for the Long run results. Further diagnostics like hetroskedasticity, Auto-correlation etc can be done by selecting view and residuals diagnostics thereafter.

1. how do you define long-run and short run equations.

2. cointegration always indicate long run equilibrium relationship between the variables.

3. three cointegration equations refers to three different dimensions of long-run realtionships between the variables.

4. can you please explain how you are using ARDL for long run and short run separately. is it based on time period of your data?

if you provide more details, i may help you further in this regard

Firstly, as Mr. Potharla informed you that you can't analyze the models in short time. Because all the coentegrioan tests analyze the relatioans between variables in long time.

Secondly, i undestood that there ise 7 variable in your analysis. Maybe you can make Cointegration Test with 2 or 3 models. And I recomend you to use diffirent cointegration tests except Johansen..

I need to examine the impact of fiscal policy on economic growth. GDP is the dependent variable, while the independent variables are Productive Government Expenditure, Non-productive Government Expenditure, DistortionaryTaxation, Non-distortionary Taxation and Foreign Aid. I would like to determine the extent to which these independent variables cause changes in GDP.