Good morning,
I have a panel data dataset containing 21 companies of the same country (X), and collecting observations for 20 years. So number of individual= 21 and T=20 with 5 independent variable, as EBITDA, liquidity and so on. I need to add to the regression a control variable (GDP per capita) of the country X, that is a time series variable with T=20.
How I have to set the variable? I am performing the analysis in STATA. I need to repeat the control variable 21 times in order to construct it( as presented below)?
company A : t=1 EBITDA(t=1) GDP country X(t=1)
company A : t=2 EBITDA(t=2) GDP country X(t=2)
[...]
company A : t=21 EBITDA(t=21) GDP country X(t=21)
company B : t=1 EBITDA(t=1) GDP country X(t=1)
company B : t=2 EBITDA(t=2) GDP country X(t=2)
[...]
such that Gdp country X (t=1) is repeted 20 times and so on for all T?
thank you so much for your help and time
You'll have to repeat the variable 21 times for each cross-section (company)
In the first place, unless you actually need that variable, it is not a good idea to introduce only country-level time-varying variables. It would be better to include time fixed effects (e.g., using year dummies). Furthermore, your strategy will force you to take into account clustering at the year level, a bad idea given that you have only 5 years.
In the secon place, in order to add the variable. I would just create a dataset (in Stata or Excel) containing the variable year ("year") and the variable GDP ("gdp"). Let's call it "aux.dta"
Then you just merge. From your original file you can type:
merge m:1 year using "aux.dta"
I hope it helps.
It is easy to include a trend variable, but the question is why and why it should be a control variable. Such a trend is only a normal exogenous variable. I am with Jose-Ignacio that - at least for the beginning you should estimate equaions without trend. You have data for 21 companies. As I assume, they are very different in size, the main problem is to standardize the data, e.g. by dividing them by one of the data, which you think is the best indicator for size (e.g. employees, turnover). Otherwise. one gets too a large weight of the data of the biggest companies.
- Badges
- Science topic