Page 1 of 1

How to confront real data

PostPosted: Sun Aug 19, 2007 3:15 am
by jiholee
Hello,

I would like to estimate a simple growth model using the attached mod file in 'Practicing Dynare'.
I wonder how I can confront the real data rather than using simulated data.
For example, the steady state of consumption 'c' is 1.4.
How can I manipulate the actual data to match this number before estimating some parameters?

Would you please giving me any advice?

PostPosted: Sun Aug 19, 2007 8:10 am
by pun_star
i doubt if these steady state values , have any meaning at all, when it comes to estimating the model. unless of course, the steady-state c is a parameter to be estimated. the great ratios like C/Y or INV/Y can be calibrated by looking at the data series. these are usually never estimated.
these models are defined in deviations off a stationary point. to estimate your model, just make sure your data-transformation is consistent with the theoretical model. i usually use the model(linear) option. so the steady state is in effect zero for all (log-linearised) variables.
i hope this makes sense.

cheers
estefan