Blanchard-Kahn conditions not holding

Stochastic simulations
1

Hi, I am trying to solve a 3 country model, with one center and two peripheries, with financial frictions a la Gertler and Karadi and with a Macroprudential tax on the rate of intermediation (This model is a simplification of Banerjee, Devereux and Lombardo (2016) but with macroprudential policies and an additional emerging country).

When estimating the decentralized model I get that the Blanchard Conditions Don’t hold:

EIGENVALUES:
     Modulus             Real        Imaginary

   2.016e-18        2.016e-18                0
   2.618e-18       -2.618e-18                0
   9.964e-18        9.964e-18                0
   2.568e-17        2.568e-17                0
   3.041e-17       -3.041e-17                0
   1.281e-15        1.281e-15                0
   2.521e-15        2.521e-15                0
   3.759e-15       -3.759e-15                0
   5.781e-15        5.781e-15                0
   1.203e-14        1.203e-14                0
   1.378e-14       -1.378e-14                0
   8.714e-14       -8.714e-14                0
   5.861e-09       -5.861e-09                0
   5.861e-09        5.861e-09                0
        0.15             0.15                0
        0.15             0.15                0
        0.15             0.15                0
        0.15             0.15                0
        0.15             0.15                0
        0.15             0.15                0
       0.775           0.7298           0.2607
       0.775           0.7298          -0.2607
      0.8021           0.7992          0.06834
      0.8021           0.7992         -0.06834
      0.8401           0.8307           0.1252
      0.8401           0.8307          -0.1252
      0.9191           0.9188           0.0229
      0.9191           0.9188          -0.0229
      0.9647           0.9647                0
      0.9771           0.9771                0
       1.067            1.067                0
        1.12            1.118          0.06263
        1.12            1.118         -0.06263
       1.131             1.13          0.04346
       1.131             1.13         -0.04346
       1.279            1.279                0
   5.432e+16        5.432e+16                0
   7.745e+16        7.745e+16                0
   1.102e+17       -1.102e+17                0
    1.22e+17        -1.22e+17                0
   4.642e+17       -4.642e+17                0
   7.518e+17        7.518e+17                0
         Inf             -Inf                0
         Inf             -Inf                0


There are 14 eigenvalue(s) larger than 1 in modulus 
for 16 forward-looking variable(s)

The rank condition ISN'T verified!

Any idea on how to solve this?. Or where in the model could be the problem?

CGMPCoord_Lv_steadystate.m (4.6 KB)
CGMPCoord_Lv.mod (11.7 KB)
CGMPCoord_sseqs.m (5.9 KB)
find_ss.m (719 Bytes)

2

This is hard to tell. It may have to do with the timing structure and the closing of the model. You are missing two unstable roots here. I can only give you the generic advice: start from a simpler working version of the model and then see when/why the problem appears.

3

Thank you so much! Johannes.

Indeed, I checked the timing and had now it’s working.