Creating leading variable

[Ruben Peijs]

Dear all,

I have a question regarding creating a leading variable.
In my data, I have some missing years from different companies and I only want to lead successive year. So I know this code for creating a lagged variable:
by state: gen lag1 = x[_n-1] if year==year[_n-1]+1

For my dataset, the code would be:

by gvkey: gen tobinsqlead = tobinsq_w[_n-1] if year==year[_n-1]+1

Only thing is, this is the code for creating a lagged variable if you have missing years, but I want to create a leading variable.
I am not able to rewrite the code into a leading one. When I "play" with the - and + I constantly get the message "weights not allowed"
Can someone help me with this?

Regards,

[Clyde Schechter]

I could show you how to fix up that code. But, why go to all this trouble? It is unlikely you really need a variable that contains the lead (or lag) of another variable in your data set.

xtset gvkey year

will lay the groundwork for using time-series operators (-help tsvarlist- for more information). Then if you need to use the lead of tobinsq in some command just write F.tobinsq where you would otherwise write the name of such a variable. Stata will do the calculation of the lead values on the fly for you without cluttering up your dataset with a new "variable" that contains no new information.

In the unlikely event you need to use leads or lags in a command that does not allow time-series operators, then you can go ahead and create your lead variable, before that command, with

gen tobinsqlead = F.tobinsq

and use that.

Nearly all Stata commands for which the use of lagged or leading values would be relevant allow time-series operators. Working with them is very easy and avoids the potential errors of "rolling your own."

[Ruben Peijs]

Thank you very much for your help!