Problem with xtabond2 command

[Roxanna Sh]

Hello,
I am trying to use xtabond2 command to run a model. My data span over four years and I have about 600,000 observations. I tried to write my model based on the following explanation
"We limited the endogeneity to IT investment and firms' related strategies and only use the second lags of first-differences as instruments "


xtabond2 firmperformance_ln l.firmperformance_ln straone stratwo strathree strafour strafive it_investment_ln
gmm (l. reven_ln straone stratwo strathree strafour strafive it_investment_ln) iv ( l2.straone l2.stratwo l2.strathree l2.strafour l2. strafive , eq(level) )
twostep robust small orthogonal

I ran the mode but I got the error of "Equation not identified. Regessors outnumber instruments." Then, I added the number of instruments as follows


xtabond2 irmperformance_ln l.firmperformance_ln straone stratwo strathree strafour strafive it_investment_ln
gmm (l. reven_ln straone stratwo strathree strafour strafive it_investment_ln) iv ( l2.straone l2.stratwo l2.strathree l2.strafour l2. strafive firmsize firmage )


xtabond2 irmperformance_ln l.firmperformance_ln straone stratwo strathree strafour strafive it_investment_ln
gmm (l. reven_ln straone stratwo strathree strafour strafive it_investment_ln) iv ( l2.straone l2.stratwo l2.strathree l2.strafour l2. strafive firmsize firmage yone ytwo ythree yfour )
twostep robust small orthogonal

And also
xtabond2 reven_ln l.reven_ln straone stratwo strathree strafour strafive it_investment_ln
gmm (l. reven_ln straone stratwo strathree strafour strafive it_investment_ln) iv ( l1.straone l1.stratwo l1.strathree l1.strafour l1. strafive firmsize firmage yone ytwo ythree yfour )
twostep robust small orthogonal

But I still got the same error. My instruments clearly are more than the regressors but I do not know why I got this error

Some points

1- Firms can pursue more than one strategy so that the strategy variable can not be regarded as a factor variable
2- My data are unbalanced so that's why used two-step system gmm



Can somebody please help me out?
Thanks

[Roxanna Sh]

Can someone help me with this error?

[Marco Astori]

I think you should reduce the number of instruments through the suboption collapse, into the gmmstyle option, or through a specific interval for lags which tells Stata which you want to exploit. Maybe Kripfganz could tell you more because I'm not an expert of the command because I'm studying it too.

[Sebastian Kripfganz]

Did you include the options in the same command line as the variable list? Stata executes the commands line by line. If you typed everything in the same way as in your post, then Stata would execute

xtabond2 firmperformance_ln l.firmperformance_ln straone stratwo strathree strafour strafive it_investment_ln

without the options that follow on the next lines. Thus, from Stata's point of view, no instruments have been specified. You need to add a comma after the above variable list and then all the options on the same line after the comma.

If you want to break the command over multiple lines in a do-file (not the command window), then you could also follow the approach that is explained in the technical note at the bottom of the following help file:

help comments

[Roxanna Sh]

Thank you so much that was my problem. Just one more question in the sentences of "We limited the endogeneity to IT investment and firms' related strategies and only use the second lags of first-differences as instruments " The second lags of first-differences as instruments mean the following model. I mean is the iv part of my model is what this sentence says.



xtabond2 firmperformance_ln l.firmperformance_ln straone stratwo strathree strafour strafive it_investment_ln ,
gmm (l. reven_ln straone stratwo strathree strafour strafive it_investment_ln) iv ( l2.straone l2.stratwo l2.strathree l2.strafour l2. strafive , eq(level) ) twostep robust small orthogonal

[Sebastian Kripfganz]

I would interpret the sentence to mean gmm(varlist, lag(2 2) eq(level)), where varlist needs to be replaced by the relevant list of variables pertaining to "IT investment and firms' related strategies".