2nd Difference Unit Root Test for Panel Least Squares

[Red Dragon]

Hi All,

I am very stuck, and hope you guys can help me

I am doing a Panel Least Squares analysis for my undergraduate Dissertation and have been told by my peers that i have to test for stationarity. I have a variable which is only stationary at the 2nd difference level however i don 't know what command i enter to create a new version of that variable. If that makes sense.

My variable is Trade and it is not stationary at the normal and 1st difference level but is at the 2nd. I know i type dtrade=d(trade) for first difference but what do i do for second difference?

Thanks for your help in advance.

[EViews Gareth]

Code: Select all

dtrade = d(trade,2)

[Red Dragon]

Thanks Gareth, that was quick.

1 more question now i have run a new regression with the 2nd difference variable and I ran the Hausman test and it shows on random that Chi squared stat is 0.016890. Which i think means i can either use fixed or random, correct? When i compare my random model against my fixed i get a D.W:1.7 for Fixed and 1.6 for Random. Now here is my dilema which one do i pick as my trade figure is showing a insignificant probability but is positively correlated on both models. What do i pick.

I have 5 Cross sections and 7 variables including my y variable. is there a way of doing random with Swamay-Arora? Because at current i am using Wallace-Hussain.