Demeaning data on a rolling window

[Giacomo Porcellotti]

Hello Statalists,

I need to dynamically demean a time series by the following backward looking rolling window of forty quarters:

That is, I need to replace each observation with the observation itself minus the mean computed in a rolling window of 10 years.

Can you help me please?

Giacomo

[George Ford]

tssmooth ma might work.

[George Ford]

clear
set obs 30
g N = _n
tsset N
tssmooth ma maN = N , window(10)
replace maN = . if N<10