Create a diagonal matrix to overcome heteroscedasticity

[JNP]

Hi,
I just started using EViews 7 and I admit I'm not very adept in econometrics, too. I tried to find an answer to my problem in the EViews user guide and here in the forum, but without succes.

I have a data series containing stock prices, 100 observations approximately. I simply want to perform a Chow test, to check whether a break has occurred at a specific date.
My equation specification is: prices c prices(-1)
I know how to conduct the Chow test, but the latter requires homoscedasticity, which unfortunately does not apply in my case. I know I can solve the problem by multiplying the original regression equation by an appropriate diagonal matrix, however I do not really know how to proceed in EViews.

Anyone able to help me out would really make my day!

[trubador]

What you are referring to is known as "weighted least squares" and there is a good discussion (along with examples) of this subject in EViews' users guide (volume II chapter 19 page 36), not to mention Startz' "EViews Illustrated" (http://www.eviews.com/illustrated/EView ... trated.pdf, chapter 14 page 335).

[JNP]

Thank you, especially for that second reference, looks quite useful!