Re: [Gretl-users] Rolling OLS

It seems to work well! After making all arrangements, my code looks like that: smpl roll_firstobs roll_firstobs+roll_window_int loop j = roll_firstobs + roll_window_int .. roll_firstobs + roll_window_int + roll_nr loop foreach i excshare string temp=sprintf("beta%d",i) if sum(missing(excshare.$i)) # This avoids calculating betas when not all data points are available loop m = $t2 .. roll_lastobs series @temp[m] = NA endloop else ols excshare.$i const excindex@ind loop m = $t2 .. roll_lastobs if ok(excshare.$i[m])==0 # This avoids having betas for dead periods series @temp[m] = NA else series @temp[m] = $coeff[2] endif endloop endif endloop if j<roll_firstobs + roll_window_int + roll_nr smpl +roll_step +roll_step endif endloop This rolling regression works really well. It loops into a list of excess returns and computes sequential OLS, keeping the main coefficient (which is beta) in a series for each stock. The only problem I still have is on the time taken when the list of stocks is large... -- Filipe Rodrigues da Costa Send me an email to: filipe(a)pobox.io Reach me through Telegram at: https://t.me/rodriguesdacosta On Fri, 27 Oct 2017, at 12:34, Sven Schreiber wrote: > Am 27.10.2017 um 13:27 schrieb Filipe Rodrigues da Costa: > >>> T = $nobs >>> >>> scalar window_size = 20 >>> scalar k = $nobs - window_size + 1 >>> series b = NA >>> >>> smpl 1 window_size >>> loop i = window_size .. T >>>     ols AMZN const SP500 >>>      if i < T >>>          smpl +1 +1 >>>      endif >>> endloop >>> >>> smpl full >> So far so good, this works quite well. But let's say the data covers 100 >> periods for SP500 but only 60 for AMZN (no data for the last 40). >> My question is as follows: Is there a simple way of imposing the routine >> to only estimate OLS when we have the full 20 data points for AMZN and >> 20 for SP500? > There are various possibilities. For example, check for any missings in > the sample like this (untested, bugs likely!): > > <hansl> > open denmark > T = $nobs > window_size = 20 > > list all = LRM LRY > loop i=window_size..T > start = i - window_size + 1 > smpl start i > if sum(missing(all)) == 0 > ols LRM const LRY > endif > > endloop > </hansl> > > hth, > sven > _______________________________________________ > Gretl-users mailing list > Gretl-users(a)lists.wfu.edu > http://lists.wfu.edu/mailman/listinfo/gretl-users _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users