Missing values

[Jordi Imbrechts]

Dear all,

I have a question about missing values. I need to create a dummy variable that gives the value "1" when there is a dismissals of the workforce (of Belgian firms, period 2011-2020) >= 0.10
Information about data:
- panel data.
- period 2011-2020
- around 185 000 observations

I have calculated the percentage of dismissals as following:

gen Dismissalstototalemployees = Dismissals/L.Totalemployees

In the next step I generate my dummy variable:

gen Collectiefontslag_10procent = 0

Because of the fact that the first observation for each firm can't be calculated, I use the following command to copy this missing value into my dummy variable:

replace Collectiefontslag_10procent = . if Dismissalstototalemployees == .

Next up I replace all the values of this dummy variable to "1" when "Dismissalstototalemployees" >= 0.10:

replace Collectiefontslag_10procent = 1 if Afdankingentovaantalwn >=0.10

Data would be this right now:

input long ID byte    Collectiefontslag    float(Afdankingentovaantalwn    Collectiefontslag_10procent)
1 .           . 1
1 0 .0042328043 0
1 0  .007556675 0
1 0  .009876544 0
1 0   .01178782 0
1 0  .008716707 0
1 0  .007302824 0
1 0  .009398496 0
1 0  .007111111 0
1 0  .004752475 0
2 .           . 1
2 0   .05988024 0
2 0  .064935066 0
2 0   .05072464 0
2 0   .01090909 0
2 0   .04477612 0
2 0  .029661017 0
2 0   .04329005 0
2 0   .01826484 0
2 0  .004464286 0
3 .           . 1
3 0  .008974708 0
3 0  .008898194 0
3 0  .007133421 0
3 0  .006935525 0
3 0  .008522036 0
3 0   .01293661 0
3 0  .005654633 0
3 0  .006094906 0
3 0  .005598622 0
4 .           . 1
4 0  .004157044 0
4 0 .0009191177 0
4 0  .002764977 0
4 0  .003678161 0
4 0  .003628118 0
4 0 .0040964955 0
4 0  .002743484 0
4 0  .002710027 0
4 0  .002604167 0
5 .           . 1
5 0 .0020686088 0
5 0 .0023098793 0
5 0 .0016697588 0
5 0 .0017384585 0
5 0 .0018552876 0
5 0   .00252419 0
5 0  .002361022 0
5 0  .002364066 0
5 0 .0010660981 0
end

"Collectiefontslag" = the dummy that I want to create in Stata, but instead I did this in Excel (to compare if the dummy generated in Stata is ok, but for robustness checks I need this code in Stata aswell)
Afdankingentovaantalwn = Dismissalstototalemployees

PROBLEM: As you can see, every time there is a missing value in "Dismissalstototalemployees" (Afdankingentovaantalwn), Collectiefontslag_10procent will show "1". As you can compare to the dummy created in Excel (Collectiefontslag), this is not what I want.

How can I solve this problem, that when there is a missing value, Stata doesn't interpret it as ">= 0.10"?

Thanks in advance,
Jordi

[Jordi Imbrechts]

Nevermind, stupid question. Fixed it. Thanks.

[Clyde Schechter]

Your post is confusing because you switch back and forth between English and another language (Afrikaans? Dutch?) in your variable name. But I think I know what you want to do, and it can be done more simply, and in one line. I will stick to English names here:

gen Collectiefontslag_10procent = (Dismissals/L.Totalemployees >= 0.10) if !missing(Dismissals, L.Totalemployees)

Added: Crossed with #2.

I will also add a general comment: in Stata, missing values are interpreted as larger than any real number. There is no way to get Stata to do otherwise, but, as here, you can use -if- conditions to exclude missing values from the calculation altogether.

[Jordi Imbrechts]

Thank you Clyde!