Am 12.11.2013 11:43, schrieb Claudio Shikida (敷田治誠 クラウジオ): > Hi > > Just one question: is there any script for gretl that calculate DOLS > (Stock & Watson 1993)? >
In principle it's easy to do, since x(+1) gives you the lead (and thus x(+1)-x the lead diff, and so on), which you can use with the ols command; if necessary, create the regressors in a loop. Nothing special that you need apart from that, I'd say. However, it's true that while there's 'lags()' to generate list of lags quickly, this doesn't seem to cover leads, and IMHO this would be a useful feature. Also, because it's impossible to enter leads via GUI into the standard OLS-dialog, perhaps it's worth thinking about a separate DOLS entry in the time series submenu? cheers, sven
