Rolling regression with de-meaned variables

[hbana]

I understand how to do the basic rolling regression program using the FOR loop.

Could someone please indicate how to conduct a regression of de-meaned values using a rolling window? I need to subtract the mean from both the dependent and independent variables. In this case the mean is the sample mean in each sub-sample (each window).

eg. y - ybar = m(x-xbar) + residual

Thanks!

[EViews Gareth]

Apart from the obvious suggestion that you simply include a constant in your regression, to literally do what you want, you simply do this:

Code: Select all

equation eq1.ls y-@mean(y) x-@mean(x)