Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: AW: Estimation with ivregress 2sls


From   "Martin Weiss" <[email protected]>
To   <[email protected]>
Subject   st: AW: Estimation with ivregress 2sls
Date   Thu, 28 May 2009 14:27:45 +0200

<> 

Try this example for a start, where the results match exactly...

*************
/*replication of -ivregress- 
results via OLS */ 

webuse hsng2, clear

ivregress 2sls rent pcturban/* 
 */ (hsngval = faminc reg2-reg4),/* 
 */ first


//replication
reg hsngval pct faminc reg2-reg4
predict pred

reg rent pred pcturban
*************



HTH
Martin

-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Christian Weiss
Gesendet: Donnerstag, 28. Mai 2009 12:38
An: [email protected]
Betreff: st: Estimation with ivregress 2sls

Dear Statalisters,

I am trying to estimate the following model with the help of the
ivregress 2sls command.


          "ivregress 2sls tobinsq capextosales debttoassets rndtosales
(cr5 = netincome5yr marketcap beta propertyrights), first"

However, if i "manualy" estimate the 2sls model by:

          "regress cr5 netincome5yr marketcap beta propertyrights
           predict cr5_est
           regress tobinsq cr5_est capextosales  debttoassets rndtosales"

I reveive slightly different results.

I am even more puzzled that when using the ivregress 2sls command  in
the First Stage regression automatically ALL used variables are
included, not just the ones specified in the brackets (allthough I
dont want that?)

Where is my mistake?

Below you find the Results of both estimations

          "ivregress 2sls tobinsq capextosales debttoassets rndtosales
(cr5 = netincome5yr marketcap beta propertyrights), first"

First-stage regressions


Number of obs   =        736
F(   7,    728) =       8.48
Prob > F        =     0.0000
R-squared       =     0.0754
Adj R-squared   =     0.0665
Root MSE        =    20.2763


cr5       Coef.   Std. Err.      t    P>t     [95% Conf. Interval]

capextosales     .115936   .0680177     1.70   0.089    -.0175982
.2494703
debttoassets   -.0465086   .0470884    -0.99   0.324    -.1389538
.0459366
rndtosales   -.2852332   .1662676    -1.72   0.087    -.6116544    .0411879
netincome5yr   -.0013242   .0004404    -3.01   0.003    -.0021887
-.0004597
marketcap   -3.62e-06   7.87e-07    -4.61   0.000    -5.17e-06   -2.08e-06
beta   -7.296444   1.905024    -3.83   0.000    -11.03644   -3.556448
propertyri~s   -.2005358   .0681679    -2.94   0.003    -.3343649
-.0667067
_cons     62.3268   5.991164    10.40   0.000     50.56478    74.08882



Instrumental variables (2SLS) regression               Number of obs =
736
Wald chi2(4)  =   71.21
Prob > chi2   =  0.0000
R-squared     =       .
Root MSE      =   .9057


tobinsq       Coef.   Std. Err.      z    P>z     [95% Conf. Interval]

cr5   -.0160655   .0060373    -2.66   0.008    -.0278984   -.0042325
capextosales    .0086722   .0030052     2.89   0.004     .0027821
.0145623
debttoassets   -.0119467   .0020976    -5.70   0.000    -.0160579
-.0078355
rndtosales    .0302816   .0076401     3.96   0.000     .0153074    .0452559
_cons    1.786574   .2357556     7.58   0.000     1.324502    2.248647

Instrumented:  cr5
Instruments:   capextosales debttoassets rndtosales netincome5yr marketcap
beta propertyrights




"regress cr5 netincome5yr marketcap beta propertyrights
           predict cr5_est
           regress tobinsq cr5_est capextosales  debttoassets rndtosales"




. regress cr5 netincome5yr marketcap beta propertyrights

      Source |       SS       df       MS              Number of obs =
1832
-------------+------------------------------           F(  4,  1827) =
41.11
       Model |   176575.35     4  44143.8376           Prob > F      =
0.0000
    Residual |  1961834.11  1827  1073.80083           R-squared     =
0.0826
-------------+------------------------------           Adj R-squared =
0.0806
       Total |  2138409.46  1831  1167.89157           Root MSE      =
32.769

----------------------------------------------------------------------------
--
         cr5 |      Coef.   Std. Err.      t    P>|t|     [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
netincome5yr |  -.0021707    .000593    -3.66   0.000    -.0033338
-.0010075
   marketcap |  -6.14e-06   1.10e-06    -5.59   0.000    -8.30e-06
-3.99e-06
        beta |   -8.37647   1.723757    -4.86   0.000    -11.75721
-4.995729
propertyri~s |  -.4999151   .0557635    -8.96   0.000    -.6092821
-.3905482
       _cons |   91.66744   4.600183    19.93   0.000     82.64527
100.6896
----------------------------------------------------------------------------
--

.
. predict cr5_est
(option xb assumed; fitted values)
(289 missing values generated)

.
. regress tobinsq cr5_est capextosales  debttoassets rndtosales

      Source |       SS       df       MS              Number of obs =
741
-------------+------------------------------           F(  4,   736) =
22.93
       Model |  65.8800181     4  16.4700045           Prob > F      =
0.0000
    Residual |  528.556382   736  .718147259           R-squared     =
0.1108
-------------+------------------------------           Adj R-squared =
0.1060
       Total |    594.4364   740  .803292433           Root MSE      =
.84744

----------------------------------------------------------------------------
--
     tobinsq |      Coef.   Std. Err.      t    P>|t|     [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
     cr5_est |  -.0142285    .003345    -4.25   0.000    -.0207954
-.0076616
capextosales |   .0071401   .0027939     2.56   0.011     .0016551
.0126251
debttoassets |  -.0112407   .0019551    -5.75   0.000     -.015079
-.0074024
  rndtosales |   .0341779   .0069125     4.94   0.000     .0206074
.0477484
       _cons |   1.760613    .148781    11.83   0.000     1.468527
2.052698
----------------------------------------------------------------------------
--
*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index