# Re: st: 2SLS: same variable as instrument and dependent variable?

 From Austin Nichols To statalist@hsphsun2.harvard.edu Subject Re: st: 2SLS: same variable as instrument and dependent variable? Date Tue, 10 Nov 2009 16:50:32 -0500

```Newer commands do not allow such a nonsensical specification:

. sysuse auto
(1978 Automobile Data)

. ivreg price (mpg=turn price)

Instrumental variables (2SLS) regression

Source |       SS       df       MS              Number of obs =      74
-------------+------------------------------           F(  1,    72) =   30.35
Model |    68760967     1    68760967           Prob > F      =  0.0000
Residual |   566304429    72  7865339.29           R-squared     =  0.1083
Total |   635065396    73  8699525.97           Root MSE      =  2804.5

------------------------------------------------------------------------------
price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
mpg |  -408.9816   74.23439    -5.51   0.000    -556.9652   -260.9981
_cons |   14875.46   1614.256     9.22   0.000      11657.5    18093.42
------------------------------------------------------------------------------
Instrumented:  mpg
Instruments:   turn price
------------------------------------------------------------------------------

. ivregress gmm price (mpg=turn price)
price specified as both regressand and excluded exogenous variable
r(498);

. ivregress liml price (mpg=turn price)
price specified as both regressand and excluded exogenous variable
r(498);

. ivregress 2sls price (mpg=turn price)
price specified as both regressand and excluded exogenous variable
r(498);

On Tue, Nov 10, 2009 at 3:59 PM, Christian Weiß <mail@cweiss.org> wrote:
> Dear Statalist,
>
> I am estimating the following 2SLS model:
>
> Firm Performance = A B (C = E F Firm Performance)
>
> I noticed that whereas ivreg allows to use the SAME variable as
> overall depdndent varialble AND instrument, ivreg2 does not allow to
> do so.
> Personally, I consider it as awkward to somehow predict a variasble by
> itself, even on a 2sls, however I am wondering what is actually
> correct?
>
> Best
> Chris
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```