Hi!

I estimated the attached model and every estimation check seems fine. Although, the std of the delta shock process (e_d) is too high in relation with the other estimated standard deviations. Specifically, the posterior mean for e_d is around 17, while the other processes posterior means are significantly lower.

Can anyone tell me if this is fine or not?

Thanks!

That depends very much on the economics of the model. You as the model builder need to decide whether such large fluctuations in the exogenous variable may be plausible.

Thanks, Johannes!

I apologize by my ignorance but, how i’m supposed to know if a shock has to have large or low std.deviation based on the model? I mean, what feature of the model should I look at to verify this?

Look at the equation

```
log(delta)=(1-rho_d)*log(bar_delta)+rho_d*log(delta(-1))+e_d;
```

So a one standard deviation shock of `e_d=17`

will increase delta roughly by a factor of 17. I don’t know what `delta`

is, but it’s often implausible.

Delta is a job separation shock, which is the same for both types of workers, high and low skilled

A 17-fold change seems strange. Then you should think hard why the model needs such large separation shocks to explain the observed data.