SB_K
December 20, 2022, 1:24pm
1
Dear all.

I set up an extended NK model with capital accumulation, calvo pricing, and so on.

I solved steady state and linearization of the model used in dynare.

But model_diafnostics continues to show
MODEL_DIAGNOSTICS: No obviable problems with this mod-file detected

However, I can continuously see following error :
There are 10 eigenvalue(s) larger than 1 in modulus for 9 forward-looking variable(s)
The rank condition ISN’T verified!

A lot of forum posts say that this is a matter of timing

I cannot find the problem of timing in my model.

Please help me solve this problem.

I am attaching my model.
prac_n2.mod (6.4 KB)

Thank you very much for your help in advance.

Your model features explosiveness. You should try to find out why.

SB_K
December 23, 2022, 6:20am
3
Thank you professor. I am going to find the explosion of model.

SB_K
December 29, 2022, 2:22pm
4
Dear professor Johannes Pfeifer.

I modified my model many point.

Now I use LAMBDA as an endogeneous variable.
(It is used usually in lagrangian)

My model can be solved without any error massage.

However, there is a non zero value in steady state (even my model is linear) that is the LAMBDA.

Always, the value is over 0.97 (almost 1).

I do not convince the result because I know it is required there are no non zero value in steady state.

I tried to remove the variable but when I removed the variable, my model do not varified the rank and cannot br solved.

Is it possible? and is there any problem in my model?
(I am attaching my model and data)

Best regard.
ALL_n2.mod (10.7 KB)
final_data_obs26_1.xls (44 KB)

Your linearization

```
c = (1/(1 + ((h^2)*beta))) * (h*beta*c(+1) + h*c(-1) - h*(1-beta) - (1-h*beta)*(1-h)*lambda - (1-h*beta)*(1-h)*price_gamma_cd + (1-h)*(1-h*beta)) ;
```

must be wrong. Why is there a constant term `(1-h)*(1-h*beta`

in there?

SB_K
January 8, 2023, 5:29am
6
Dear professor Johannes Pfeifer,

Thank you for your comment.

I modified my model and remove some irregular parts in my model (ex: lambda and some constant term you mentioned).

My model can be estimated now.

Thank you very much professor.

Best regard,