singularity problem

[dean0504]

Hi All! When I run the attached model, the BK condition is not satisfied. And I run the model_diagnostics(M_,options_,oo_) commend, it says there is a colinear relationship between Equation 3 and 37. However, equation 3 is the Euler equation and 37 stands for the exogenous interest rate. If following the suggestion of Dynare, I do not know which equation I need to delete and which equation I lost. Please help! Thanks!

[jpfeifer]

You have a unit root in your model. In that case, there will always be collinearity. Check whether the unit root is intended. If not, there is already the source of your problem.

[dean0504]

Thanks Jpfeifer! Is the unit root caused by the values of the parameters?

[jpfeifer]

I don't know. Often it is, but it could also be that you implicitly set an unspecified parameter like

Code: Select all

z=z(-1)+eps_z

Here the AR1-process has a coefficient of one, but is not parameterized.

In short: it could be a parameter, as mistake in an equation or a feature of your model.

[dean0504]

Thank you so much, Johannes! I found two equations with unit roots which are

Code: Select all

B_s*B(-1)+r_s*B_s*(r(-1)+B(-1))+Y_s*P_s*(Y+P)+W_s*W=G_s*P_s*(G+P)+B_s*B

and

Code: Select all

P_s*C_s*(P+C)+B_s*B=(1+tauw)*W_s*L_s*(W+L)+B_s*(1+r_s)*B(-1)+r_s*B_s*r(-1)

.
Because I set the steady state value of interest rate r_s equals to 1/beta-1, the unit root occurs. Would you please tell me in this case, what should I do? Recode B(-1) to B and B to B(+1)?

[jpfeifer]

I don't know, because it is your model. It might be the problem discussed in http://ideas.repec.org/a/eee/inecon/v61y2003i1p163-185.html