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

# Antwort: st: Hausman Query

 From Justina Fischer <[email protected]> To [email protected] Subject Antwort: st: Hausman Query Date Wed, 5 Jan 2011 11:22:17 +0100

Hi

I am not an expert on this (any around ?), but I would just do what Stata suggests.
Hausman needs to calculate "[(V_b-V_B)^(-1)]", which, obviously, cannot be done for one of your estimators.

My guess would be that House price causes that problem, or any variable with an extremely small coefficient estimate. Try rescaling....

Justina
[email protected] schrieb: -----

An: "[email protected]" <[email protected]>
Von: "Ross, Andrew" <[email protected]>
Gesendet von: [email protected]
Datum: 05.01.2011 09:50AM
Thema: st: Hausman Query

Hello

I am currently using Stata to investigate new firm formation for 32 Scottish regions over a 10 year period. I was wondering, if you might offer your opinion on a Stata related matter, that I have encountered?

As you can see from the attached output I've run a fixed and random effects model and followed this by the hausman test. However, the hausman highlights a 'note', as you will see from the attached output.

I was wondering, if you could shed any light on this note and what is means? I have asked four other people and they do not know. One suggested it may be a result of over parameterisation given the small size of the panel, but they are not sure.

Many thanks.

Andrew

xtreg Lab_TP Wge_grow Pop_grow Log_unemployed NVQ4_pop House_price LQ_agric LQ_man LQ_b
> s Pop_density Gov_sector Small_bus, fe vce (robust)

Fixed-effects (within) regression               Number of obs      =       320
Group variable: Region                          Number of groups   =        32

R-sq:  within  = 0.2061                         Obs per group: min =        10
between = 0.0019                                        avg =      10.0
overall = 0.0030                                        max =        10

F(11,277)          =      3.10
corr(u_i, Xb)  = -0.9507                        Prob > F           =    0.0006

(Std. Err. adjusted for clustering on Region)
------------------------------------------------------------------------------

|               Robust
Lab_TP |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------
-------------+------
Wge_grow |  -.0265289   .0270548    -0.98   0.328     -.079788    .0267303
Pop_grow |   .4054536   .5845519     0.69   0.489    -.7452749    1.556182
Log_unempl~d |   4.109563   4.222209     0.97   0.331     -4.20213    12.42126
NVQ4_pop |  -.0295052   .0431429    -0.68   0.495    -.1144348    .0554243
House_price |   .0000483   .0000121     3.99   0.000     .0000245    .0000722
LQ_agric |   .1687096   .4963323     0.34   0.734    -.8083529    1.145772
LQ_man |  -.1523606   .4695697    -0.32   0.746    -1.076739    .7720178
LQ_bs |   2.506032   1.647939     1.52   0.129    -.7380422    5.750107
Pop_density |   .0236083   .0110802     2.13   0.034     .0017962    .0454204
Gov_sector |  -.0001643   .0309392    -0.01   0.996    -.0610701    .0607415
Small_bus |   .3920981   .4383467     0.89   0.372     -.470816    1.255012
_cons |   -32.9382   42.48215    -0.78   0.439    -116.5671    50.69067
-------------+----------------------------------------------------------
-------------+------
sigma_u |  17.794849
sigma_e |   2.771809
rho |   .9763121   (fraction of variance due to u_i)
------------------------------------------------------------------------------

. estimates store fixed

. xtreg Lab_TP Wge_grow Pop_grow Log_unemployed NVQ4_pop House_price LQ_agric LQ_man LQ
> _bs Pop_density Gov_sector Small_bus, re vce (robust)

Random-effects GLS regression                   Number of obs      =       320
Group variable: Region                          Number of groups   =        32

R-sq:  within  = 0.1667                         Obs per group: min =        10
between = 0.6091                                        avg =      10.0
overall = 0.4986                                        max =        10

Random effects u_i ~ Gaussian                   Wald chi2(12)      =   3152.08
corr(u_i, X)       = 0 (assumed)                Prob > chi2        =    0.0000

(Std. Err. adjusted for clustering on Region)
------------------------------------------------------------------------------
|               Robust
Lab_TP |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------
-------------+------
Wge_grow |  -.0254291   .0298964    -0.85   0.395     -.084025    .0331667
Pop_grow |   .8175778   .5211693     1.57   0.117    -.2038953    1.839051
Log_unempl~d |  -4.643811   2.993296    -1.55   0.121    -10.51056    1.222941
NVQ4_pop |    .008423   .0484623     0.17   0.862    -.0865613    .1034074
House_price |   .0000265   .0000108     2.46   0.014     5.39e-06    .0000476
LQ_agric |   .7014816   .4126435     1.70   0.089    -.1072848    1.510248
LQ_man |   .2703715   .5041589     0.54   0.592    -.7177617    1.258505
LQ_bs |   4.506697   1.720704     2.62   0.009      1.13418    7.879214
Pop_density |   .0023373   .0006452     3.62   0.000     .0010727    .0036018
Gov_sector |  -.0345817    .034529    -1.00   0.317    -.1022574    .0330939
Small_bus |   .6680098   .1846674     3.62   0.000     .3060683    1.029951
_cons |   -42.7439   17.37493    -2.46   0.014    -76.79813   -8.689662
-------------+----------------------------------------------------------
-------------+------
sigma_u |  2.2647466
sigma_e |   2.771809
rho |  .40033378   (fraction of variance due to u_i)
------------------------------------------------------------------------------

. estimates store random

. hausman fixed random

Note: the rank of the differenced variance matrix (10) does not equal the number of
coefficients being tested (11); be sure this is what you expect, or there may
be problems computing the test.  Examine the output of your estimators for
anything unexpected and possibly consider scaling your variables so that the
coefficients are on a similar scale.
---- Coefficients ----
|      (b)          (B)            (b-B)     sqrt(diag(V_b-V_B))
|     fixed        random       Difference          S.E.
-------------+----------------------------------------------------------
-------------+------
Wge_grow |   -.0265289    -.0254291       -.0010997               .
Pop_grow |    .4054536     .8175778       -.4121243        .2647328
Log_unempl~d |    4.109563    -4.643811        8.753374        2.977789
NVQ4_pop |   -.0295052      .008423       -.0379283               .
House_price |    .0000483     .0000265        .0000218        5.53e-06
LQ_agric |    .1687096     .7014816       -.5327721        .2758099
LQ_man |   -.1523606     .2703715       -.4227322               .
LQ_bs |    2.506032     4.506697       -2.000664               .
Pop_density |    .0236083     .0023373         .021271        .0110614
Gov_sector |   -.0001643    -.0345817        .0344174               .
Small_bus |    .3920981     .6680098       -.2759117        .3975498
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test:  Ho:  difference in coefficients not systematic
chi2(10) = (b-B)'[(V_b-V_B)^(-1)](b-B)
=        5.58
Prob>chi2 =      0.8493
(V_b-V_B is not positive definite)

Andrew Ross
PhD Candidate
School of Accounting, Economics & Statistics
Craiglockhart Campus
Room 1/38
EH14 IDJ
Email: [email protected]

Edinburgh Napier University is Edinburgh's top university for graduate employability (HESA 2010), and proud winner of the Queen's Anniversary Prize for Higher and Further Education 2009, awarded for innovative housing construction for environmental benefit and quality of life.

This message is intended for the addressee(s) only
and should not be read, copied or disclosed to anyone else outwith the University without the permission of the sender. It is your responsibility to ensure that this message and any attachments are scanned for viruses or other defects.
Edinburgh Napier University does not accept liability for any loss or
damage which may result from this email or any attachment, or for errors or omissions arising after it was sent. Email is not a secure medium. Email entering the University's system is subject to routine monitoring and filtering by the University.

Edinburgh Napier University is a registered Scottish
charity.
Registration number SC018373

*
*   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/