Page 1 of 1

Negative R squared

Posted: Sun Nov 01, 2015 10:31 am
by midobrown
Good day dears,
Your kind support please to help me with the following problem when analyzing the data on EViews:
I got negative R squared.
knowing that, I am using dummy variables to check the existence of seasonality during the week is stock market index returns which used in natural logarithm.
the equation does not contains any constant and in using GARCH model.

Thanks for your great efforts.

Regards,

Re: Negative R squared

Posted: Sun Nov 01, 2015 12:06 pm
by EViews Gareth
There is nothing to say that R-squared should be positive when using a GARCH model with no constant.

Re: Negative R squared

Posted: Sun Nov 01, 2015 12:14 pm
by startz
However, you may have a sign of misspecification. If you run the equation OLS rather than GARCH, what happens?

Re: Negative R squared

Posted: Sun Nov 01, 2015 11:47 pm
by trubador