I believe there is a bug in the asreg function. When I run rolling window regressions on panel data (firm * quarter), the predictions and residuals are identical in each quarter. E.g. when I run...
asreg y x, by(qtr) fitted window(qtr 4)
... all predicted values in a particular quarter are identical, and so are all residuals.

This problem does *not* occur when I do not specify a window, ie when running
asreg y x, by(qtr) fitted

Something seems be wrong with the "window" option.

Has anybody had the same issue?