sspace results do not make sense

[EviewsUser1]

I am running a fairly simple state space model but the smoothed state smsv4 is behaving bizarrely. The model does not converge using the default algorithm but even when I switch to eviews legacy I get bizarre results for smsv4. Workfile is attached.

series gap_res=gap-1.27*gap(-1)+.333*gap(-2)
series r_lag=r(-1)+r(-2)
vector(5) svec0
svec0.fill 0,0,0,0,0

':::::::: set up Kozicki statespace model ::::::::::::::


smpl 1960 2015q2
sspace Kozicki

Kozicki.append @signal gap=sv1+c(5)*(r_lag-sv4-sv5)
Kozicki.append @state sv1=c(1)*sv1(-1)+c(2)*sv2(-1)+[ename=e1,var=@exp(c(3))]
Kozicki.append @state sv2=sv1(-1)
Kozicki.append @state sv3=sv3(-1)+[ename=e2,var=@exp(c(4))]
Kozicki.append @state sv4=sv3(-1)
Kozicki.append @state sv5=sv4(-1)
Kozicki.append @evar cov(e1,e2)=0
Kozicki.append @mprior svec0

Kozicki.ml(maxit=1000)

Kozicki.makestates(t=smooth) sm*

qgraph r smsv4
stop