How to solve:Log data density [Laplace approximation] is NaN

[Syscus]

I don’t know why the error “Log data density [Laplace approximation] is NaN. ”is reported. I tried many methods and it seems to be a problem with the model setting, but this is just the simplest New Keynesian model.
test.mod (3.8 KB)
test_data.m (3.1 KB)

[jpfeifer]

You fix h=1/3; in steady state, but why should that be true for any value of alpha you estimate?

[Syscus]

First of all, thank you very much for your reply. I tried to estimate other parameters. There was no problem when I estimated the technology shock and its persistence coefficient, but the same problem occurred when I estimated other parameters, such as government purchase shock and monetary policy shock. Exogenous labor supply seems to be a common method for solving steady-state problems. As a beginner who uses Bayesian estimation for the first time, I really find it difficult to find a solution.

[jpfeifer]

See

Steady-state negative Frisch Elasticity

The Frisch elasticity is still a parameter. I don’t understand what you are doing. When you calibrate labor in steady state, the disutility parameter needs to be adjusted accordingly. But that is different from the Frisch elasticity. See e.g. which features a unit Frisch elasticity, but sets the labor disutility parameter phi to have labor as 1/3

Anything that affects the steady state will have such a problem.