Unusual Problem

I’m trying to implement a model with exogenous bond-market segmentation. There are 3 types of agents who can buy only one kind of bond and one unrestricted agent. There are 3 kinds of perpetuites with different duaration plus a short-term bond. The estrange thing is that the model doesn’t work. I get the message: There are 36 eigenvalue(s) larger than 1 in modulus
for 39 forward-looking variable(s)

The rank conditions ISN’T verified!"

However, when I delete the variables ksi_r1(+1), ksi_r2(+1) and ksi_r3(+1) from the equations of the lines 193,194 e 195, the model just work perfectly.

Can anybody help me to understand that?

The file is attached.
Model4.mod (13.3 KB)

the equation of ksi_r1,ksi_r2 and ksi_r3, you have two sets of this three functions, why?

It’s a segmented market (4 bonds types and 4 agents) where one agent is unrestricted and the others can only buy 1 kind of bond.

Does the model then still have full rank if it is linearized (no covariances to pin down the optimal allocation)?

When I eliminate the ksi`s it does have full rank. But without that, no.