Dear all,

There seems to be a bug in dsge_likelihood.m that occurs when calling
estimation after identification. In the attached mod-file in
dsge_likelihood.m we have in lines 628 following:

    if analytic_derivation==1,

analytic_deriv_info={analytic_derivation,DT,DYss,DOm,DH,DP,asy_Hess};
    else

analytic_deriv_info={analytic_derivation,DT,DYss,DOm,DH,DP,D2T,D2Yss,D2Om,D2
H,D2P};
        clear DT DYss DOm DH DP D2T D2Yss D2Om D2H D2P,
    end

that is if analytic_derivation==2 set by identification we delete DH. But in
line 714 we access DH, which was deleted.

                if analytic_derivation,
                    for j=1:pp,
                        tmp(j,:)=DH(j,j,:);
                    end
                    DH=tmp;
                end

It seems we have to account for this case.

Best,

Johannes
--
Johannes Pfeifer
Haußerstr. 29
72076 Tübingen
Tel.: +49-(0)7071-6396184
Mobil.: +49-(0)170-6936820
Germany