14. Panel Unit Root Testing using EViews || Dr. Dhaval Maheta

[@bandanamondal664]

Sir should we have to confirm the data is stationary only for the dependent variable or in case of all the independent and dependent variables for further proceeding to the regression?

[@ananyabhatia5843]

Very well Explained Sir. but my question is if one of the test shows that the series is stationary at level and other test states that series is stationary at first difference then what which test we will consider then ?

[@DhavalSaifaleeAaryash]

Majority wins

[@ananyabhatia5843]

@@DhavalSaifaleeAaryash but objective of my study is to compare the performance of banks before and after pandemic. Then which test of stationarity I will go for ( or I will check the result of all the test). Moreover what if the variables came out to be stationary on first difference then how to proceed further ?

[@aminaahmedalibelal5676]

well and easy explained Dr. so if my data is I(1) does that means i will use 1st difference variable instead of original variable in my analysis?? Thank youi

[@DhavalSaifaleeAaryash]

Yes

[@shakshisharma9839]

Sir I have a doubt regarding stationarity test and Normality test. The data I am working on is only for a singly year with 6 variables of 100 companies. then which test I need to apply?. ADF is for time series where for cross section I am unable to find.

[@DhavalSaifaleeAaryash]

You can run simple regression. Making one variable dependent and other independent

[@shakshisharma9839]

@@DhavalSaifaleeAaryash yes sir I have done that but normally before running regression don't we have to check stationarity and Normality?

[@shakshisharma9839]

I have 2 dependent variables and it both resulted insignificant results. I applied ADF and after first level of difference I somehow managed to get significant results of only one DV.

[@DhavalSaifaleeAaryash]

@@shakshisharma9839 stationary for variables, normality for residuals