Dear Saqib

You should test for unit roots and cointegration tests even if the econometric model is based on an economic model. This is necessary as non-stationary data will give you spurious regressions unless the variables are cointegrated.

Thanks.

Shafiullah

Dear Saqib,

The condition of testing the stationarity is essential as it directs to the appropriate method of estimating the coefficients of the model. A spurious regression can be dangerous for making inferences. 

Best Wishes

Muhammad Shafiullah & Muhammad Atif

Thank you so much for your suggestions. But i have seen many articles in world-class journals when they deal with Gravity they do not really care about Unit root testing or mentioning anywhere in research. That's the reason i have raised a question here.