Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# RE: st: why reg3 dropped constant term ?

 From DE SOUZA Eric <[email protected]> To "[email protected]" <[email protected]> Subject RE: st: why reg3 dropped constant term ? Date Sun, 2 Jan 2011 13:49:55 +0100

```The only reason that I can see is perfect collinearity between the explanatory variables.

Eric de Souza
College of Europe
Brugge (Bruges), Belgium
http://www.coleurope.eu

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of KAMAL MANJU
Sent: 02 January 2011 04:27
To: statalist-for help STATA
Subject: RE: st: why reg3 dropped constant term ?

Dear sir/Dr./Prof,
Happy New Year to all.

Sorry for being late  to respond to your request for clarification on my question. I was not online the past few days, and because I got few automatic responses from some members that says that they are out of office since until after new year break,so I also guessed that many list members may have gone for leave too.

Anyway, thanks to all, and thank you specially Austin, for the reply and the request for clarification.

Here is what I have typed in Stata and the Stata estimated output.

Manually (using hand to type in Stata command window,

reg3 (Leverage: d.lvr=l.lvr d.rwar roa llpta lta liq hhi y_2 y_3 y_4 y_5 y_6 y_7 y_8) (Prisk: d.rwar = l.rwar d.lvr llpta liq lta hhi y_2 y_3 y_4 y_5 y _6 y_7 y_8), small

Note that:
D.cbf and D.rwar are the only two endogenous variables in this model, and that all other variables are   considered to be exogenous in this model.

y_2 and y_8 refer to year 2001 and year 2007 respectively s based on my year dummy coding

Using Stata menu I also selected the same variables in the same arrangement but the output was the same.

The sample output below is from Stata estimation of this model.

===============Stata output==============================================

. reg3 (Leverage:d.lvr=l.lvr d.rwar roa llpta lta liq  hhi y_2 y_3 y_4 y_5 y_6 y_7 y_8) (Prisk: d.rwar = l.rwar d.lvr llpta liq lta  hhi y_2 y_3 y_4 y_5 y
> _6 y_7 y_8),small

Three-stage least-squares regression
----------------------------------------------------------------------
Equation          Obs  Parms        RMSE    "R-sq"     F-Stat        P
----------------------------------------------------------------------
Leverage          152     13    3.226453    0.3097       7.69   0.0000 Prisk             152     12    .0663633    0.3523       9.18   0.0000
----------------------------------------------------------------------

------------------------------------------------------------------------------
|      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------
-------------+------
Leverage     |
lvr |
L1. |  -.3381578   .0438067    -7.72   0.000    -.4243969   -.2519186
rwar |
D1. |   19.31579   5.935037     3.25   0.001     7.631911    30.99967
roa |   .4260782   .2626352     1.62   0.106    -.0909527    .9431091
llpta |   4.980128   56.11149     0.09   0.929    -105.4825    115.4428
lta |  -1.534593   .2895698    -5.30   0.000    -2.104648   -.9645376
liq |     .77008   1.208503     0.64   0.525    -1.609013    3.149173
hhi |   1.255695   .2308078     5.44   0.000     .8013206     1.71007
y_2 |   2.398955   1.032184     2.32   0.021     .3669695    4.430941
y_3 |   4.346167   .9869655     4.40   0.000     2.403199    6.289135
y_4 |   4.321573   1.082539     3.99   0.000     2.190458    6.452689
y_5 |   3.466324   1.099247     3.15   0.002     1.302315    5.630333
y_6 |   2.999611   1.097867     2.73   0.007     .8383201    5.160903
y_7 |   2.601068   1.043284     2.49   0.013     .5472302    4.654905
y_8 |  (dropped)
_cons |  (dropped)
-------------+----------------------------------------------------------
-------------+------
Prisk        |
rwar |
L1. |  -.4186713   .0569779    -7.35   0.000    -.5308395    -.306503
lvr |
D1. |   .0073145   .0024066     3.04   0.003     .0025769    .0120521
llpta |   .0715227   .9750749     0.07   0.942    -1.848037    1.991082
liq |   .2557745   .0346649     7.38   0.000     .1875322    .3240168
lta |    .014243   .0046491     3.06   0.002     .0050907    .0233953
hhi |  -.0062491    .003048    -2.05   0.041    -.0122494   -.0002487
y_2 |   .0418541   .0196936     2.13   0.034     .0030847    .0806234
y_3 |   .0097648   .0199548     0.49   0.625    -.0295188    .0490483
y_4 |   .0014469   .0187886     0.08   0.939    -.0355409    .0384348
y_5 |   .0274229   .0189197     1.45   0.148    -.0098228    .0646687
y_6 |   .0215651   .0193331     1.12   0.266    -.0164946    .0596249
y_7 |   .0191237   .0196772     0.97   0.332    -.0196133    .0578608
y_8 |  (dropped)
_cons |  (dropped)
------------------------------------------------------------------------------
Endogenous variables:  D.lvr D.rwar
Exogenous variables:   L.lvr roa llpta lta liq hhi y_2 y_3 y_4 y_5 y_6 y_7
y_8 L.rwar
------------------------------------------------------------------------------

========================End of Stata output==========================================
.

----------------------------------------
> Date: Tue, 28 Dec 2010 18:38:18 -0500
> Subject: Re: st: why reg3 dropped constant term ?
> From: [email protected]
> To: [email protected]
>
> KAMAL MANJU :
> Can you report what you typed and what Stata estimated?
> It seems unlikely the constant was actually dropped...
>
> On Fri, Dec 24, 2010 at 12:34 PM, KAMAL MANJU  wrote:
> >
> > i sent this before but bounced
> >
> > Dear sir/Profs/Dr. and fellow students
> >
> > I am trying to estimate a simultaneous equation model for a panel of
> > firms from four different countries. I am running the analysis
> > separately by country= by country,so I have 4 separate data sets but all have exactly the same variables and the model will be applied to all constituency. My data set is annual and spans for 8years period.
> >
> > I need to estimate two equations simultaneously (the same equation to be repeated for each country case).
> >
> > I need to include the constant term as well as year dummies (to be consistent with the literature in this area).
> >
> > However for unknown reasons stata reg3 always drop the constant term
> > from the two equations in some countries whenever I included the full year dummies(8year dummies -1year=2C base year) in the model.
> >
> > I need your kind advice on this problem. I mean how can I make sure that the constant term is reported not drop from the model output.
> >
> > Alternatively, I would like to seek advice from experts on this issue or whoever have encountered similar problem how to go about.
> >
> > Is this a statistical/econometric problem if I cannot report the
> > constant term and include it in my analysis?  Any guide to relevant rea reachable resources is highly appreciated.
> >
> > I have already spent quite a time searching on stata frequently
> > asked questions and consulted other sources but to no informative answer. I really stuck at this point in my analysis I need help.
> >
>
> *
> * 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/

*
*   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/
```