Use of Time dummy in Panel Data

[Neeraj Kumar]

Hi,
I am using five year panel data for my research. While going through with review of literature. I have found that some paper include time dummies and state dummies in there analysis and some does not. They gave reason for including time dummies is It allows to control for time-specific fixed effects. Sorry if the question sound like basic question but I just want to know is it necessary to include time dummies and state dummies while using panel data.

[Mario Ferri]

Hi,
I am using five year panel data for my research. While going through with review of literature. I have found that some paper include time dummies and state dummies in there analysis and some does not. They gave reason for including time dummies is It allows to control for time-specific fixed effects. Sorry if the question sound like basic question but I just want to know is it necessary to include time dummies and state dummies while using panel data.

Theory and your research assumptions will give you the answer. Grab a good textbook for a start!

[Carlo Lazzaro]

Neeraj:
in addition to Mario's wise advice, it is usual to include -i.timevar- in -fe- panel data regression to investigate the effect of time on within panel varuatiin of the dependent variable, when adjuted for the other predictors.
That said, another issue seems to creep up from your question: the role of time-invariant variable that differ from -panelid- in -fe- panel data regression. As they' do not change across time, they are wiped out by the -fe- estimator and, therefore, no coefficient is estimated for them.