Statalist


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

Re: st: re: overid error


From   John Antonakis <john.antonakis@unil.ch>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: re: overid error
Date   Mon, 23 Nov 2009 19:27:08 +0100

Dear Kit:

-overid- works like a charm now. Thank you very much for providing Stata users with such a great resource and for updating the code!

There is a small discrepancy in the overidentification statistic.

. reg3 (dv = iv) (iv = ex1 ex2), 2sls

Two-stage least-squares regression
----------------------------------------------------------------------
Equation          Obs  Parms        RMSE    "R-sq"     F-Stat        P
----------------------------------------------------------------------
dv              10000      1    1.074345    0.2886    1008.56   0.0000
iv              10000      2     1.27832    0.4436    3985.68   0.0000
----------------------------------------------------------------------

------------------------------------------------------------------------------
| Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dv           |
iv | .2989434 .0094132 31.76 0.000 .2804927 .3173941 _cons | .0047041 .0107439 0.44 0.662 -.0163548 .025763
-------------+----------------------------------------------------------------
iv           |
ex1 | .796612 .0126401 63.02 0.000 .7718364 .8213877 ex2 | .8169581 .0128188 63.73 0.000 .7918322 .842084 _cons | -.0174155 .0127855 -1.36 0.173 -.0424762 .0076452
------------------------------------------------------------------------------
Endogenous variables:  dv iv
Exogenous variables:   ex1 ex2
------------------------------------------------------------------------------

. overid

Number of equations : 2
Total number of exogenous variables in system : 3
Number of estimated coefficients : 5
Hansen-Sargan overidentification statistic :     0.820
Under H0, distributed as Chi-sq(1), pval = 0.3653

What is a little odd, though, is that the Hansen Sargan statistic returned by overid (using a simple two-equation model with 2SLS) is not the same as ivreg2:

. ivreg2 dv (iv=ex1 ex2),

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics consistent for homoskedasticity only

Number of obs = 10000 F( 1, 9998) = 1008.56 Prob > F = 0.0000 Total (centered) SS = 16221.61478 Centered R2 = 0.2886 Total (uncentered) SS = 16221.64239 Uncentered R2 = 0.2886 Residual SS = 11539.86191 Root MSE = 1.074

------------------------------------------------------------------------------
dv | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
iv | .2989434 .0094123 31.76 0.000 .2804957 .3173911 _cons | .0047041 .0107428 0.44 0.661 -.0163514 .0257596
------------------------------------------------------------------------------
Underidentification test (Anderson canon. corr. LM statistic): 4436.329 Chi-sq(2) P-val = 0.0000
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic): 3985.676 Stock-Yogo weak ID test critical values: 10% maximal IV size 19.93 15% maximal IV size 11.59 20% maximal IV size 8.75 25% maximal IV size 7.25
Source: Stock-Yogo (2005).  Reproduced by permission.
------------------------------------------------------------------------------
Sargan statistic (overidentification test of all instruments): 0.768 Chi-sq(1) P-val = 0.3808
------------------------------------------------------------------------------
Instrumented:         iv
Excluded instruments: ex1 ex2
------------------------------------------------------------------------------

Best,
John.
____________________________________________________

Prof. John Antonakis
Associate Dean Faculty of Business and Economics
University of Lausanne
Internef #618
CH-1015 Lausanne-Dorigny
Switzerland

Tel ++41 (0)21 692-3438
Fax ++41 (0)21 692-3305

Faculty page:
http://www.hec.unil.ch/people/jantonakis

Personal page:
http://www.hec.unil.ch/jantonakis
____________________________________________________



On 23.11.2009 17:10, Kit Baum wrote:
> <>
> Martin said
>
> " We have determined that this is caused by a bug in -ivtobit- return
> values, which we have worked around and reported to StataCorp."
>
>
> So the update to -overid- is somewhat ephemeral itself then? Once the
> official -ivtobit- has been fixed, -overid- can revert to its state of last
> night?
>
> No. For one thing the reg3 handling code has been updated. For the ivtobit issue, the problem -- as Martin reported -- was that ivtobit is returning two or more variable names in e(depvar). We have pulled off word 1 of depvar, which will work even when e(depvar) is corrected to be a single word.
>
> Kit Baum | Boston College Economics & DIW Berlin | http://ideas.repec.org/e/pba1.html > An Introduction to Stata Programming | http://www.stata-press.com/books/isp.html > An Introduction to Modern Econometrics Using Stata | http://www.stata-press.com/books/imeus.html
>
>
> *
> *   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index