I run a simple model with no problem, but when I consider the consumption inertia in one sector, then I got the error:

There are 11 eigenvalue(s) larger than 1 in modulus
for 12 forward-looking variable(s)
The rank condition ISN’T verified!

I only change one sector of the model: ln( c_t) ln( c_t - eta c_{t-1}), the model just adds an extra Lagrange operator lamda_t, and I got the problem. It is always a timing error, so when I change evey lamda_t to lamda_{t-1}, the model runs. However, the impulse response is totally different from my benchmark. I think something wrong, but I don’t know how to solve it.

I would appreciate someone could help me! Thank you very much in advance.

The code is followed, I only change the real sector. thy6_1.mod (7.0 KB)

Dr.jpfeifer,thank you a lot. When I take eta = 0 then I get the original model, but when I increase the number of eta , when it > ~ 0.4 then the model gets the error, however, I have to calibrate eta to 0.4, so what may cause the problem ?

That very much sounds like a parameterization issue. Have you tried changing the fiscal and monetary policy rules to see whether that affects the result.

Actually, I used a Taylor Rule, but I also met another the same error: I can not set the parameter of interest sticking to inflation \rho_\pi to 1.5, which is common in the paper, I have to set it to ~0.7, then the model runs, so I turn to this monetary ruler, then I can not set the \eta.