Syntax to include time trend in panel data analysis

[Mathilde Bauwin]

Hi everyone,

I have a rather simple question but just want to be clear about it:

I want to run a panel data analysis and I will try several commands to do it as I have time-invariant variables (xthtaylor, xtabond2 and I will try xtdpdml, which is not yet official though, as far as I know).

In these various models, I want to estimate the effect of the time trend, which means I would like to include a continuous variable for time (t), and a quadratic one as well (t²).

Can I include such time variables in these models, and if it is the case, should I use xtset like this:

xtset id t

or like this:

xtset id

Thank you very much

[Christos Makridis]

If you have a variable called year, just add in year as an explanatory variable and/or it's square or whatever polynomial. And, if you work with quarterly data, you'd want to do "egen tid=group(year quarter)" and then do the same thing but instead with tid.