# st: lag endog variables as instruments in ivreg2

 From mameezah <[email protected]> To [email protected] Subject st: lag endog variables as instruments in ivreg2 Date Sun, 28 Jun 2009 20:06:12 +1000

```Dear Statalisters,
Possibly my question is more of an econometric than something to do
with using Stata. However, running out of my other means I thought to
use Stata listers to clarify some of my confusion.

I am estimating my model with 2sls using ivreg2 in Stata and have used
lag values of my endogenous variables as instruments. I am a bit
concern of Anderson canon. corr. LR statistic  of 5378.702. It would
be helpful to have your valuable comment on this statistics and also
using lag of endogenous variable/s as instrument in 2sls.

In this regard, I have included the stata output of my estimated model

Mamiza Haq
University of Queensland

Below is the output:
ivreg2 r7 v33 v37 v40 v43 m15   y97 y98 y99 y00 y01 y02 y03 y04 y05
y06 (v29 v45
>  = l.v29  l.v45 v25),  endog(v29 v45) robust
IV (2SLS) regression with robust standard errors
------------------------------------------------
Number of obs =     3952
F( 17,  3934) =    53.13
Prob > F      =   0.0000
Total (centered) SS     =  585.3429765                Centered R2   =   0.2912
Total (uncentered) SS   =  683.5763175                Uncentered R2 =   0.3931
Residual SS             =  414.8672882                Root MSE      =     .324
------------------------------------------------------------------------------
|               Robust
r7 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
v29 |   .1501156   .0241436     6.22   0.000      .102795    .1974361
v45 |   .2575129   .0681073     3.78   0.000     .1240252    .3910007
v33 |   .0207883   .0105656     1.97   0.049     .0000801    .0414965
v37 |  -.0962742   .0085038   -11.32   0.000    -.1129414    -.079607
v40 |  -.0169613   .0020569    -8.25   0.000    -.0209927     -.01293
v43 |  -.2082114   .0455637    -4.57   0.000    -.2975146   -.1189082
m15 |   .0183414   .0008159    22.48   0.000     .0167423    .0199404
y97 |   .0363958   .0215968     1.69   0.092    -.0059332    .0787247
y98 |   .0892018   .0261002     3.42   0.001     .0380464    .1403572
y99 |   .0253146   .0213949     1.18   0.237    -.0166186    .0672478
y00 |   .0049437   .0190883     0.26   0.796    -.0324687    .0423561
y01 |   .0129248   .0197953     0.65   0.514    -.0258733     .051723
y02 |   .0272125   .0197594     1.38   0.168    -.0115152    .0659401
y03 |   .0449799    .019871     2.26   0.024     .0060334    .0839264
y04 |   .1097093   .0217121     5.05   0.000     .0671543    .1522643
y05 |   .1857581   .0252958     7.34   0.000     .1361793     .235337
y06 |   .1453426   .0238486     6.09   0.000     .0986003    .1920849
_cons |  -1.026569   .0678847   -15.12   0.000     -1.15962   -.8935171
------------------------------------------------------------------------------
Anderson canon. corr. LR statistic (identification/IV relevance test):5378.702
Chi-sq(2) P-val =    0.0000
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         2.998
Chi-sq(1) P-val =    0.0834
-endog- option:
Endogeneity test of endogenous regressors:                               7.722
Chi-sq(2) P-val =    0.0210
Regressors tested:    v29 v45
------------------------------------------------------------------------------
Instrumented:         v29 v45
Included instruments: v33 v37 v40 v43 m15 y97 y98 y99 y00 y01 y02 y03 y04 y05
y06
Excluded instruments: L.v29 L.v45 v25
------------------------------------------------------------------------------
. test y97 y98 y99 y00 y01 y02 y03 y04 y05 y06
( 1)  y97 = 0
( 2)  y98 = 0
( 3)  y99 = 0
( 4)  y00 = 0
( 5)  y01 = 0
( 6)  y02 = 0
( 7)  y03 = 0
( 8)  y04 = 0
( 9)  y05 = 0
(10)  y06 = 0
chi2( 10) =  135.27
Prob > chi2 =    0.0000
*
*   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/
```