Having troubles with ivreg2

[Franco Mariuzzo]

I have installed the newest version of ivreg2 (version 4.1.10 9Feb2016) and have run a code that was running fine with a previous version. The code now does not run and I get the error message "variable endo not found". I cannot find a solution online.

Any suggestion?

Thanks

Franco

[Marcos Almeida]

Hello Franco,

Welcome to the Stata Forum/Statalist.

Please provide the "exact" command and the data (or at least a toy example) , as recommended in the FAQ.

You may use the CODE delimiters or install the SSC dataex.

[Franco Mariuzzo]

Below is th ecommand that I use. There are two endogenous variables L1.gdl and Dver (the model is expressed in first differences):

xi: ivreg2 D1.(gdl inapp free Nidfp1 L1.rat) i.t (D1.(L1.gdl Dver) = ///
D2.ownDver D2.sprod D2.L1.owngdl L2.(gdl Dver) D1.L1.agever D1.L1.ver) if platform==0, gmm2s cluster(idf) ///
savefprefix(first)

[Marcos Almeida]

Hello Franco,

Today, I feel to be under a guessing mood. Considering I don't have much expertise with ivreg2, here are my suggestions:

You may perhaps avoid using the xi: prefix. Logically, and following the recommendations written in the FAQ, I assume you are using Stata 14.2.
Maybe you are copying and pasting in the the Command Window a code directly taken from a do-file. If so, you may wish to edit the command line accordingly.
It may well be the option "endog", instead of "endo".
I see ".t" before a variable, and that strikes me as a potential mistype.
Please re-check the correctness of the command overall. Here, I share the help files.

Hopefully that helps.

P.S.: as kindly suggested in #2, please use the CODE delimiters or install the SSC datex so as to share command and results. Thanks.

[Franco Mariuzzo]

Dear Marcos,

Thanks so much for your help. I know that xi: is no longer necessary but I used in the past and remained in some of my do files. I had no problem (and neither my coauhtors) in running that code with previous versions of Stata. I have now (as you correctly pointed out) downloaded Stata 14.2 and find that ivreg2 is no longer working. The command i.t only indicates time dummies (which can beinlcuded also in a first difference model).

The error message is "variable endo not found".

Best wishes,

Franco