Dynare OBC - Warning messages

Dynare contributions and examples
1

Hi everyone,

I am using DynareOBC to simulate a model with ZLB. However, I get this warning message.

Warning: Error running the Lemke LCP algorithm: Invalid MEX-file
‘/Users/vahagndavtyan/Desktop/dynareOBCRelease/Extern/tbxmanager/toolboxes/lcp/1.0.3/maci64/source/lcp.mexmaci64’:
dlopen(/Users/vahagndavtyan/Desktop/dynareOBCRelease/Extern/tbxmanager/toolboxes/lcp/1.0.3/maci64/source/lcp.mexmaci64, 6): Library not
loaded: @loader_path/libmwblas.dylib
Referenced from: /Users/vahagndavtyan/Desktop/dynareOBCRelease/Extern/tbxmanager/toolboxes/lcp/1.0.3/maci64/source/lcp.mexmaci64
Reason: image not found

In SolveBoundsProblem (line 123)
In FastIRFs>FastIRFsInternal (line 223)
In FastIRFs (line 36)
In dynareOBCCore (line 681)
In dynareOBCSetup (line 297)
In dynareOBC (line 711)

Should I do something to avoid this message?

2

How did you install the program? And what is your operating system. @cfp may be able to help.

3

Thank You for Your prompt reply Professor.
I simply add DynareOBC folder to my matlab path.

4

But you are on a mac, right?

5

Yes

6

I’m afraid I don’t have a Mac to help debug this. This seems to be an issue within an external component. The tbxmanager is supposed to install all required dependencies for the packages it manages (in this case lcp), so you could plausibly raise a support ticket with them.

As a work around, you can try commenting out lines 109 to 148 of Core/InnerProblem/SolveBoundsProblem.m. You may also need to call DynareOBC with the MaxParametricSolutionDimension=0 command line option. This ought to prevent it calling that LCP package.

7

Thank You for help. Now I am not getting any warning messages, but the results are unchanged.

8

I have also a question regarding the Taylor Rule. I want to have such kind of rule
R - R_ss = max(1,rho_r*(R(-1)-R_ss)+(1-rho_r) * (mu_pi * (pi-STEADY_STATE(pi))+ mu_y*(y - STEADY_STATE(y)))+u_r);

I want to impose ZLB on my rule (the gross interest rate can’t be below 1) but I get -1 in steady state, which I think comes from the max operator. How can I solve this problem?

9

That rule does not make much sense. The quantity inside the max(1,…) has steady state zero, so your equation just states that R-R_ss=1.

10

So, how should I write my interest rate rule to have both persistence and ZLB? The model is not linear.