Re: [Gretl-users] Asymmetric ECM with Gretl, is it possible ?

On Thu, 5 Apr 2012, Hanne holopainen wrote: Maybe the following example script may help: <hansl> open AWM.gdt --quiet series infl = 400*ldiff(HICP) # known beta ----------------------------- series z1 = STN - infl z1 -= mean(z1) series z2 = LTN - infl z2 -= mean(z2) list A_ECM = null loop foreach i z1 z2 series $ip = ($i>0) ? $i : 0 series $im = ($i<=0) ? $i : 0 list A_ECM += $ip(-1) $im(-1) end loop list X = infl STN LTN list DX = diff(X) list DLAGS = lags(1,DX) system name ="AsyVECM" equation d_infl const DLAGS A_ECM equation d_STN const DLAGS A_ECM equation d_LTN const DLAGS A_ECM end system estimate "AsyVECM" method=ols # unknown beta --------------------------- ols d_STN const DLAGS --quiet gam = $coeff scalar a11 = 0 scalar a12 = 0 scalar a21 = 0 scalar a22 = 0 scalar b1 = 1 scalar b2 = 1 nls d_STN = gam[1] + lincomb(DLAGS, gam[2:]) + ec(-1) series ec = 0 series z1 = STN - b1 * infl series z2 = STN - b2 * LTN series ec = (z1>0) ? a11 * z1 : a12 * z1 series ec += (z2>0) ? a21 * z2 : a22 * z2 params gam a11 a12 a21 a22 b1 b2 end nls </hansl> Don't be too picky about the economics or the econometrics in the example, it's just to show you how you can use hansl to do a few things that you may find useful. :) Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti