Re: [Gretl-users] how to reproduce Verbeek's Table 5.4 (one-step GMM)?

On Tue, 15 Nov 2011, Qi Shi wrote: There are a few problems with your script, but the most important one is that once you have defined a series as a function of a parameter, the contents of that series don't change just because you change the parameter's value; that is the way a spreadsheet works, not a programming language like hansl. For example: after <hansl> scalar delta=0.5 scalar gamma=0.5 series e0=delta*cons^(-gamma)*(1+rf)-1 </hansl> the contents of e0 are set to e0 1959:02 -0.4993965 1959:03 -0.5020569 1959:04 -0.4966915 1959:05 -0.5009903 1959:06 -0.4981722 1959:07 -0.4983170 1959:08 -0.4984926 1959:09 -0.4996040 1959:10 -0.4966996 1959:11 -0.4982402 ... Then, the parameters delta and gamma are iteratively changed within the gmm block, but that doesn't change the contents of e0 (and e1, e2 etc, for that matter). To replicate Verbeek's example, take a look at the script below: <hansl> function series mrate(series cons, scalar discrat, scalar riskad) return discrat * cons ^ (-riskad) end function open pricing.gdt --quiet set force_hc on # not terribly correct, but # necessary to replicate exactly # build the instruments list exlist = null loop i=1..10 --quiet series ex$i = r$i - rf exlist += ex$i end loop series gf = 1 + rf scalar d = 0.5 scalar g = 10 matrix V = I(11) series m oc1 # initialize the series with the # orthogonality conditions /* first step */ gmm series m = mrate(cons,d,g) series oc1 = m * gf - 1 orthog oc1 ; const orthog m ; exlist weights V params d g end gmm /* iterated GMM */ gmm series m = mrate(cons,d,g) series oc1 = m * gf - 1 orthog m ; exlist orthog oc1 ; const weights V params d g end gmm --iterate </hansl> Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche r.lucchetti(a)univpm.it http://www.econ.univpm.it/lucchetti