Day of the Week Effect with Dummy Variables, Garch(1,1)

[Mar93]

Ok thank you startz. And you'd first try it with all five dummies and then afterwards with the constant and four dummies, only if the first one doesn't work? so are both ways equal or is one of them preferable?

[startz]

Both ways give the same answer, just different interpretations.

[Mar93]

So when I put c, d2, d3, d4 and d5 in my model, does c represent monday? Wouldn't it be the average of all days? and the p-values of d2-d5, do they tell me whether the difference to te constant is significant or whether the difference to zero is significant? Thank you very much.

[startz]

So when I put c, d2, d3, d4 and d5 in my model, does c represent monday? Wouldn't it be the average of all days? and the p-values of d2-d5, do they tell me whether the difference to te constant is significant or whether the difference to zero is significant? Thank you very much.

The constant gives you the intercept for Monday. The other dummies say how the intercept on other days is different from Monday's intercept.

[Mar93]

Ok thank you. So the coefficient of the constant is the mean return on Monday (Just to make sure, we're talking about c in the mean equation right?). And the p-value tells me whether the coefficient is statistically significant different from zero.
One last question: the p-values of d2-d5, do they tell me whether the difference to the constant/monday mean return is statistically signicant, or whether the difference to zero is statistically signifcant? Thank you

[startz]

The constant is the mean for Monday if there are nothing but day of week dummies in the equation, but not otherwise.

[Mar93]

ok. in my equation I only have the dummies for the days of the week. no other dummies. In the window for the garch model I just put "r c d2 d3 d4 d5".
And what about the p-values of d2-d5? do they tell me the significance of the difference to the constant, or the difference to zero? (see my last post)

[Mar93]

I would really appreciate an answer to my question about the p-values. thanks heaps. :)