# st: How to read dummy coefficients in GARCH model

 From Katia Bobulova To statalist@hsphsun2.harvard.edu Subject st: How to read dummy coefficients in GARCH model Date Sun, 30 Aug 2009 17:39:13 +0200

```Dear all,

I performed the following GARCH(1,1) model, with three dummy
variables. I included just two dummies to avoid the dummy trap. The
problem is that I am not able to read the coefficients. I mean, I know
that the coefficient of the constant is the coefficient of the omitted
variable and that the others must be read as deviation from this.
However, when I tied to do the same omitting a different dummy I could
not obtain the same coefficients, so I think that I am doing something
wrong. Without knowing this I cannot conclude my work.

ARCH family regression -- multiplicative heteroskedasticity

Sample: 3 - 2461                                   Number of obs   =      2459
Distribution: Gaussian                             Wald chi2(1)    =     82.97
Log likelihood =  8929.199                         Prob > chi2     =    0.0000

------------------------------------------------------------------------------
|                 OPG
|      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
lreturn      |
lreturn |
L1. |   .1848528   .0202941     9.11   0.000     .1450771    .2246285
_cons |    .000332    .000122     2.72   0.006      .000093    .0005711
-------------+----------------------------------------------------------------
HET          |
d_D2 |  -1.614848   .4480782    -3.60   0.000    -2.493065   -.7366312
d_D3 |  -1.294629   .4497271    -2.88   0.004    -2.176077   -.4131798
_cons |  -12.32181   .4604613   -26.76   0.000     -13.2243   -11.41933
-------------+----------------------------------------------------------------
ARCH         |
arch |
L1. |   .0800911   .0065127    12.30   0.000     .0673264    .0928558
garch |
L1. |   .8980415   .0074378   120.74   0.000     .8834637    .9126193
-----------------------------------------------------------------------------