Fixed effects panel data

[JOHAN JOHNYS]

Good evening. I want to apply fixed effects in my panel data. I have data for companies for a period of several years, and variables such as sales, Etr ratio, when they went public, etc. The variable of interest will be a dummy=0 before IPO and 1 when firm goes public and the outcome is the Etr index. I want to ask what is the code to make the panel data, to create the variable of interest I mentioned and run the fixed effects model. Please help, if you can tell me the do file (commads)so to run the commands in stata

[Carlo Lazzaro]

Johan:
as per FAQ, this is not the way this forum works.
That said, see -xtreg,fe-, -xtreg- entry, Stata .pdf manual.

[George Ford]

Is the data panel, or do you want to make it panel?

[JOHAN JOHNYS]

Is the data panel, or do you want to make it panel?

the data is a panel. I did it with xtset. I don't know how to create the variable of interest I mentioned. I have a variable called listed and it is binary. 1 if the company is listed on the stock exchange and zero if it is not. I have found the commands that I will use them but I don't know if they are correct.I thought of the commands below. To
create variable of interest I mentioned above
replace listed=0 if missing (listed)
bysort taxid:drop if listed [_n]==listed [_n-1] & listed==0
xtreg Etr sales gross profit assets,fe
What do you say;

[George Ford]

1. if listed is binary, then why are you modifying it. Or, is it 1 and missing? If so, first line is fine.
2. not sure why you are dropping things.

xtset id year
xtreg Etr listed sales gross profit assets, fe vce(cluster id)

I suspect a selection bias problem, however. Becoming listed is not random or exogenous.

[JOHAN JOHNYS]

1. if listed is binary, then why are you modifying it. Or, is it 1 and missing? If so, first line is fine.
2. not sure why you are dropping things.

xtset id year
xtreg Etr listed sales gross profit assets, fe vce(cluster id)

I suspect a selection bias problem, however. Becoming listed is not random or exogenous.

Good morning and thank you very much for the answer. I want to keep only the listed companies, i.e. the companies that all these years that I have the data have listed =0 should be removed from the sample.

[George Ford]

Ok. I thought you wanted to quantify effects of being listed or not.

[JOHAN JOHNYS]

↵

Ok. I thought you wanted to quantify effects of being listed or not.

Can you tell me about that,if it is ok
by sort company_id: egen ever_listed = max(listed)


keep if ever_listed == 1

drop ever_listed

To create the interesting variable I mentioned to you, what do you think;