Posterior estimation

Dear professor,
I am trying to estimate a NLNKM, these are the equations for capital accumulation and investment. Basically it looks like Phi_I and stderr epsMUS are in conflict with each other.
image
image

If I take the common priors in the literature, Phi_I looks ok but epsMUS does not
stderr epsMUS, 4.96, 0.01, 15, INV_GAMMA_PDF, 0.1,2;
phi_I, 1.5, 0.5, 10, NORMAL_PDF, 4, 1.5;

If I try for example these other values:
stderr epsMUS, 4.96, 0.01, 15, INV_GAMMA_PDF, 2,2;
phi_I, 1.5, 0.5, 10, NORMAL_PDF, 4, 1.5;
SE_epsMUS is super high

if I take these other values here:
stderr epsMUS, 2, 0.01, 15, INV_GAMMA_PDF, 2,2;
phi_I, 1.5, 0.5, 10, NORMAL_PDF, 1.5, 0.5;
SE_epsMUS is still high (but less than before)

Do you have any suggestions on how to deal with that? I know that the posteriors in general don’t look super good but it’s only because i have little number of iterations

You should try to understand why your model has these extreme values. rhoA looks almost like a unit root. Have you inspected whether the data matching to the model makes sense?