Creating lags of many variables all at once

[Mike Kraft]

How can I efficiently create new variables for lags of many variables in my dataset?
Say for example that I have the variables: cons inv out ae exp rev int oph lpt , and I do not want to create lags for each one separately (such as gen lag_cons=lag.cons, etc.) because it will take time especially if I have to do that for nearly 20 variables...

I tried :

local X cons inv out ae exp rev int  oph lpt
gen lag`X'=lag.`X'

But I get error message that lag operator is invalid.

[FernandoRios]

Hi Mike
First. the operator is not "lag" but simply "l." or "L.". see "help tsvarlist"
Second, why do you need to create the lags? Stata may do a better job tracking that when estimating your model.
Third. still need to create the variables one by one, but you can use a loop.
HTH
Fernando

[Nick Cox]

Stata is correct. The operator is L. See

help varlist

I am surprised that you didn't get another error message, as the left-hand side of your second command is illegal even before generate gets to look at lag.

If you really need new variables, you need an explicit loop.

local X cons inv out ae exp rev int  oph lpt  
foreach x of local X {      
    gen lag`x' = l.`x'
}

More commonly Stata just lets you specify lags on the fly when it is necessary. For example,

. webuse grunfeld, clear

. tsset
       panel variable:  company (strongly balanced)
        time variable:  year, 1935 to 1954
                delta:  1 year

. regress invest L.invest year

[Mike Kraft]

Thanks Both.
Sorry the lag in my code was obviously an typo from me. Sorry for that.
Also, yes I understand that using L. will be fine but I do need the new lag variables to perform other tests that need them to be defined.
Thanks a lot Nick, I will use this loop.
Regards