Am 25.05.2010 14:20, schrieb Claudio Shikida (敷田治誠 クラウジオ): > Hello to all, > > Straight to the point. I converted an old Eviews file to Gretl. The > series were ok, all was going well. They are time series. I updated the > series and generated the logs, d(logs) and also logs(-k) and dlogs(-k). > All of this was fine. So I went to the Instrumental Variables in order > to estimate a Bewley version of a cointegration relationship. It means I > have to use logs(-k) as instruments. Here is the problem: Gretl doesn´t > show the logs(-k), neither dlogs(-k) as options to put into the model. > So I tried to created them in this menu. New failure. > > Shortly: instrumental variables just shows me the original variables and > when I try to enter the lags of several of them as instruments, I can´t. > I presume I can change the name of the variables and solve this, but I > would like to know, first, if I am doing some foolish stuff. >
I'm not sure I understand, because for me the dialog window for IV regression clearly allows to specify lags (hit the "lags..." button at the bottom). -sven