Dynare forums

I have the following question - I added to my model monetary policy rule and it turns out that when I change the specification of this rule Blanchard Kahn conditions are either satisfied or not satisfied. The same happens when I modify Parameters in my Taylor rule.
Does this mean that there is something generally wrong with my model? I mean is it possible that another Taylor rule specification leads to model indeterminancy?

(when adding monetary authority I changed debt contracts from real terms to nominal terms)

So my question is whether the appearing model indeterminancy can be solved just by adjusting monetary rule parameters or should I search for other possible mistakes in the model? (model_diagnostics doesn't detect anything)

That is hard to tell, but generally, New Keynesian models are sensitive to different policy rules. This becomes particularly important when you have fiscal policy as there might be issues with active/passive fiscal/monetary policy (See Leeper's work). Checking for mistakes is always a good idea, but what you describe might well be an endogenous model feature.

Thank you for your reply! Now I decided to try to use optimal simple rule to find optimal monetary policy Parameters. However it doesn't work - I get either an error message "At least one of the initial parameter values for osr_params is NaN" or that "the steadystate file didn't compute steady state".

But before, with my Parameters steadystate could be computed. I used the following commands for osr:

optim_weights;
y 1;
pi 1;
y, pi 0.5;
end;

//% Initial values for parameters in Taylor rule
rR = 0.8;
rY = 0.5;
rpi = 1.5;

osr_params rpi;

osr;

steady(nocheck);

I'm confused because monetary policy rule with parameters set by me does work, however, osr cannot be computed. Why could this happen?

Here is the mod-file I use (without osr). The response to monetary policy shock looks strange - for some reason house prices increase on Impact following a contractionary monetary policy shock. The other Impulse response functions make sense for me.
Could you please take a look on my code?

Sorry, but I cannot check your model for correctness. You have to do the debugging. As always, try to simplify the model as far as possible and try to work from there. For example, start with a Taylor rule without smoothing and response to output movements.

Regarding the OSR, it seems to have been an issue with your starting values. It runs on my machine without problems.