Beginner: Generalised Schur (QZ) Decomposition Failed

Stochastic simulations
1

Dear Professor,

When running my mod file, I could not perform a stochastic simulation.

My model is a standard RBC model augmented with pandemic SIR dynamics. Derivations and MATLAB files

WORKING.mod (1.4 KB)

Project_Proposal_Presentation.pdf (271.5 KB)

are attached.

Thank you for the help!

Kind Regards,
Henry

2

model_diagnostics indicates that the following equations are collinear:

Collinear equations
1 2 3 4 5 6 9 10

3

Dear Professor,

Thank you for clarifying.

Kind Regards,
Henry

4

Dear Professor,

I’ve made some revisions to my model and I now have an indeterminacy problem. Model diagnostics says my first equations are colinear, but this is by design.

My agents face an uncertainty problem about whether they will be hospitalised in the future. The probability of becoming infected is equal to the share of the population currently hospitalised. Could the steady state hospitalisations being zero be my problem?

Kind Regards,
Henry

5

The determinacy properties from Blanchard/Kahn are local properties. It may indeed be the case that in steady state the model is indeterminate if the model is nonlinear. But then a linear approximation to the model will not work.

6

Dear jpfeifer,

If there are non-linearity issues, might Junior Maih’s RISE toolbox be a suitable tool to use?

Kind Regards,
Henry

7

It depends on the setup you have in mind. RISE’s strength is regime-switching, which is probably not what you want to do in your model.