I'm applying the HP filter to time series data to remove the cyclical component of a time series (TS) from raw data. Data do have a yearly frequency. I write in R:
lambda=100 #penalty term
cycle= hpfilter(v1,freq=lambda)$cycle
where v1 is my TS data. In this way, I get the cyclical component.
Do you agree with me that is common practice to set lambda=100 when you deal with annual data?
Yes, usually we use 100 for yearly data. See "tips" section in (https://www.mathworks.com/help/econ/hpfilter.html).