Thursday, March 10, 2011

Conditional heteroskedasticity

Finding:

There is a strong relationship between the regression error variance and the regression independent variables.

ε2 ~ Xi


A regression exhibits conditional heteroskedasticity if the variance of the regression errors (1)are not constant and (2)are related to the regression independent variables. This finding indicates that the regression exhibits conditional heteroskedasticity.

0 comments:

Post a Comment