Either use first differences or mean differences. Then run as Ordinary Least squares with clustered (Huber- White) standard errors. Best wishes. Nagarajan

I suggest using Fixed effects with robust or cluster standatd errors.

Best wishes!

Since it says FE are appropriate, proceed and use White's standard errors as suggested for better hypothesis testing...

Thanks all..

I wish to share that the FE results turns few major variables insignificant.  Any alternate solutions?

Thanks all.