Page 1 of 1

Heteroskedasticity (Need Help)

Posted: Sat Feb 07, 2015 2:48 am
by b1lal
Dear All,

I am stuck with a problem. I have made a model of growth of gdp as dependent variable and growth agriculture, services and manufacturing sectors as independent variables.

The regression was found significant along and autocorrelation is also fine. However, I am having problem in heteroskedasticity.

The heterskedasticity was found present in the model. My problem is that the data has some negative values due to which I can't take the log of the values to transform and run the model again.

Any suggestions?

Regards,
Bilal Ahmed

Re: Heteroskedasticity (Need Help)

Posted: Sat Feb 07, 2015 7:57 am
by startz
Use heteroskedastic-robust standard errors, the (h) option for least squares.